Merge mwl253 -> 10.0-base

mysql-test/r/mysqld--help.result

@@ -208,6 +208,13 @@ The following options may be given as the first argument:
  use a distinct domain_id. For simple tree-shaped
  replication topologies, it can be left at its default, 0.
  -?, --help          Display this help and exit.
+ --histogram-size=#  Number of bytes used for a histogram. If set to 0, no
+ histograms are created by ANALYZE.
+ --histogram-type=name 
+ Specifies type of the histograms created by ANALYZE.
+ Possible values are: SINGLE_PREC_HB - single precision
+ height-balanced, DOUBLE_PREC_HB - double precision
+ height-balanced.
  --ignore-builtin-innodb 
  Disable initialization of builtin InnoDB plugin
  --ignore-db-dirs=name 
@@ -506,6 +513,21 @@ The following options may be given as the first argument:
  partial_match_table_scan, semijoin, semijoin_with_cache,
  subquery_cache, table_elimination, extended_keys,
  exists_to_in } and val is one of {on, off, default}
+ --optimizer-use-condition-selectivity=# 
+ Controls selectivity of which conditions the optimizer
+ takes into account to calculate cardinality of a partial
+ join when it searches for the best execution plan
+ Meaning: 1 - use selectivity of index backed range
+ conditions to calculate the cardinality of a partial join
+ if the last joined table is accessed by full table scan
+ or an index scan, 2 - use selectivity of index backed
+ range conditions to calculate the cardinality of a
+ partial join in any case, 3 - additionally always use
+ selectivity of range conditions that are not backed by
+ any index to calculate the cardinality of a partial join,
+ 4 - use histograms to calculate selectivity of range
+ conditions that are not backed by any index to calculate
+ the cardinality of a partial join.
  --performance-schema 
  Enable the performance schema.
  --performance-schema-events-waits-history-long-size=# 
@@ -940,6 +962,8 @@ general-log FALSE
 group-concat-max-len 1024
 gtid-domain-id 0
 help TRUE
+histogram-size 0
+histogram-type SINGLE_PREC_HB
 ignore-builtin-innodb FALSE
 ignore-db-dirs 
 init-connect 
@@ -1030,6 +1054,7 @@ old-style-user-limits FALSE
 optimizer-prune-level 1
 optimizer-search-depth 62
 optimizer-switch index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,index_merge_sort_intersection=off,engine_condition_pushdown=off,index_condition_pushdown=on,derived_merge=on,derived_with_keys=on,firstmatch=on,loosescan=on,materialization=on,in_to_exists=on,semijoin=on,partial_match_rowid_merge=on,partial_match_table_scan=on,subquery_cache=on,mrr=off,mrr_cost_based=off,mrr_sort_keys=off,outer_join_with_cache=on,semijoin_with_cache=on,join_cache_incremental=on,join_cache_hashed=on,join_cache_bka=on,optimize_join_buffer_size=off,table_elimination=on
+optimizer-use-condition-selectivity 1
 performance-schema FALSE
 performance-schema-events-waits-history-long-size 10000
 performance-schema-events-waits-history-size 10

mysql-test/r/statistics.result

@@ -64,13 +64,13 @@ SELECT * FROM mysql.table_stats;
 db_name	table_name	cardinality
 test	t1	40
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -87,8 +87,8 @@ COUNT(*)
 40
 SELECT * FROM mysql.column_stats
 WHERE db_name='test' AND table_name='t1' AND column_name='a';
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
 SELECT MIN(t1.a), MAX(t1.a), 
 (SELECT COUNT(*) FROM t1 WHERE t1.b IS NULL) /
 (SELECT COUNT(*) FROM t1) AS "NULLS_RATIO(t1.a)",
@@ -99,8 +99,8 @@ MIN(t1.a)	MAX(t1.a)	NULLS_RATIO(t1.a)	AVG_FREQUENCY(t1.a)
 0	49	0.2000	1.0000
 SELECT * FROM mysql.column_stats
 WHERE db_name='test' AND table_name='t1' AND column_name='b';
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
 SELECT MIN(t1.b), MAX(t1.b), 
 (SELECT COUNT(*) FROM t1 WHERE t1.b IS NULL) /
 (SELECT COUNT(*) FROM t1) AS "NULLS_RATIO(t1.b)",
@@ -111,8 +111,8 @@ MIN(t1.b)	MAX(t1.b)	NULLS_RATIO(t1.b)	AVG_FREQUENCY(t1.b)
 vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	6.4000
 SELECT * FROM mysql.column_stats 
 WHERE db_name='test' AND table_name='t1' AND column_name='c';
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
 SELECT MIN(t1.c), MAX(t1.c), 
 (SELECT COUNT(*) FROM t1 WHERE t1.c IS NULL) /
 (SELECT COUNT(*) FROM t1) AS "NULLS_RATIO(t1.c)",
@@ -123,8 +123,8 @@ MIN(t1.c)	MAX(t1.c)	NULLS_RATIO(t1.c)	AVG_FREQUENCY(t1.c)
 aaaa	dddddddd	0.1250	7.0000
 SELECT * FROM mysql.column_stats
 WHERE db_name='test' AND table_name='t1' AND column_name='d';
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
 SELECT MIN(t1.d), MAX(t1.d), 
 (SELECT COUNT(*) FROM t1 WHERE t1.d IS NULL) /
 (SELECT COUNT(*) FROM t1) AS "NULLS_RATIO(t1.d)",
@@ -135,8 +135,8 @@ MIN(t1.d)	MAX(t1.d)	NULLS_RATIO(t1.d)	AVG_FREQUENCY(t1.d)
 1989-03-12	1999-07-23	0.1500	8.5000
 SELECT * FROM mysql.column_stats
 WHERE db_name='test' AND table_name='t1' AND column_name='e';
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
 SELECT MIN(t1.e), MAX(t1.e), 
 (SELECT COUNT(*) FROM t1 WHERE t1.e IS NULL) /
 (SELECT COUNT(*) FROM t1) AS "NULLS_RATIO(t1.e)",
@@ -206,6 +206,47 @@ WHERE t1.e IS NOT NULL AND t1.b IS NOT NULL AND t1.d IS NOT NULL)
 AS 'ARITY 3';
 ARITY 1	ARITY 2	ARITY 3
 6.2000	1.6875	1.1304
+DELETE FROM mysql.column_stats;
+set histogram_size=4;
+ANALYZE TABLE t1;
+Table	Op	Msg_type	Msg_text
+test.t1	analyze	status	Table is already up to date
+SELECT db_name, table_name, column_name,
+min_value, max_value,
+nulls_ratio, avg_frequency,
+hist_size, hist_type, HEX(histogram)
+FROM mysql.column_stats;
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_frequency	hist_size	hist_type	HEX(histogram)
+test	t1	a	0	49	0.0000	1.0000	4	SINGLE_PREC_HB	2E62A1D0
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	6.4000	4	SINGLE_PREC_HB	003FBFFF
+test	t1	c	aaaa	dddddddd	0.1250	7.0000	4	SINGLE_PREC_HB	0055AAFF
+test	t1	d	1989-03-12	1999-07-23	0.1500	8.5000	4	SINGLE_PREC_HB	009393FF
+test	t1	e	0.01	0.112	0.2250	6.2000	4	SINGLE_PREC_HB	000564E1
+test	t1	f	1	5	0.2000	6.4000	4	SINGLE_PREC_HB	3F7FBFBF
+DELETE FROM mysql.column_stats;
+set histogram_size=8;
+set histogram_type='DOUBLE_PREC_HB';
+ANALYZE TABLE t1;
+Table	Op	Msg_type	Msg_text
+test.t1	analyze	status	Table is already up to date
+SELECT db_name, table_name, column_name,
+min_value, max_value,
+nulls_ratio, avg_frequency,
+hist_size, hist_type, HEX(histogram)
+FROM mysql.column_stats;
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_frequency	hist_size	hist_type	HEX(histogram)
+test	t1	a	0	49	0.0000	1.0000	8	DOUBLE_PREC_HB	052F4363F4A1F9D0
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	6.4000	8	DOUBLE_PREC_HB	0000FF3FFFBFFFFF
+test	t1	c	aaaa	dddddddd	0.1250	7.0000	8	DOUBLE_PREC_HB	00005555AAAAFFFF
+test	t1	d	1989-03-12	1999-07-23	0.1500	8.5000	8	DOUBLE_PREC_HB	000026942694FFFF
+test	t1	e	0.01	0.112	0.2250	6.2000	8	DOUBLE_PREC_HB	000005056464E1E1
+test	t1	f	1	5	0.2000	6.4000	8	DOUBLE_PREC_HB	FF3FFF7FFFBFFFBF
+DELETE FROM mysql.column_stats;
+set histogram_size= 0;
+set histogram_type=default;
+ANALYZE TABLE t1;
+Table	Op	Msg_type	Msg_text
+test.t1	analyze	status	Table is already up to date
 CREATE TABLE t3 (
 a int NOT NULL PRIMARY KEY,
 b varchar(32),
@@ -238,16 +279,16 @@ db_name	table_name	cardinality
 test	t1	40
 test	t3	17
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
-test	t3	a	0	38	0.0000	4.0000	1.0000
-test	t3	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.1765	18.0714	2.8000
-test	t3	c	aaaa	dddddddd	0.1176	6.4000	3.7500
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
+test	t3	a	0	38	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t3	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.1765	18.0714	2.8000	0	NULL	NULL
+test	t3	c	aaaa	dddddddd	0.1176	6.4000	3.7500	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -267,16 +308,16 @@ db_name	table_name	cardinality
 test	s1	40
 test	t3	17
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	s1	a	0	49	0.0000	4.0000	1.0000
-test	s1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	s1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	s1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	s1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	s1	f	1	5	0.2000	1.0000	6.4000
-test	t3	a	0	38	0.0000	4.0000	1.0000
-test	t3	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.1765	18.0714	2.8000
-test	t3	c	aaaa	dddddddd	0.1176	6.4000	3.7500
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	s1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	s1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	s1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	s1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	s1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	s1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
+test	t3	a	0	38	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t3	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.1765	18.0714	2.8000	0	NULL	NULL
+test	t3	c	aaaa	dddddddd	0.1176	6.4000	3.7500	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	s1	PRIMARY	1	1.0000
@@ -296,16 +337,16 @@ db_name	table_name	cardinality
 test	t1	40
 test	t3	17
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
-test	t3	a	0	38	0.0000	4.0000	1.0000
-test	t3	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.1765	18.0714	2.8000
-test	t3	c	aaaa	dddddddd	0.1176	6.4000	3.7500
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
+test	t3	a	0	38	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t3	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.1765	18.0714	2.8000	0	NULL	NULL
+test	t3	c	aaaa	dddddddd	0.1176	6.4000	3.7500	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -324,13 +365,13 @@ SELECT * FROM mysql.table_stats;
 db_name	table_name	cardinality
 test	t1	40
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -365,13 +406,13 @@ t1	CREATE TABLE `t1` (
   KEY `idx4` (`y`,`x`,`d`)
 ) ENGINE=MyISAM DEFAULT CHARSET=latin1
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	x	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	y	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	x	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	y	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 ALTER TABLE t1 CHANGE COLUMN x b varchar(32), 
 CHANGE COLUMN y e double;
 SHOW CREATE TABLE t1;
@@ -390,13 +431,13 @@ t1	CREATE TABLE `t1` (
   KEY `idx4` (`e`,`b`,`d`)
 ) ENGINE=MyISAM DEFAULT CHARSET=latin1
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 ALTER TABLE t1 RENAME TO s1, CHANGE COLUMN b x varchar(32);
 SHOW CREATE TABLE s1;
 Table	Create Table
@@ -417,13 +458,13 @@ SELECT * FROM mysql.table_stats;
 db_name	table_name	cardinality
 test	s1	40
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	s1	a	0	49	0.0000	4.0000	1.0000
-test	s1	x	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	s1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	s1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	s1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	s1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	s1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	s1	x	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	s1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	s1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	s1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	s1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	s1	PRIMARY	1	1.0000
@@ -455,13 +496,13 @@ SELECT * FROM mysql.table_stats;
 db_name	table_name	cardinality
 test	t1	40
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -490,12 +531,12 @@ t1	CREATE TABLE `t1` (
   KEY `idx4` (`e`,`x`,`d`)
 ) ENGINE=MyISAM DEFAULT CHARSET=latin1
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -519,12 +560,12 @@ t1	CREATE TABLE `t1` (
   KEY `idx4` (`e`,`b`,`d`)
 ) ENGINE=MyISAM DEFAULT CHARSET=latin1
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -535,13 +576,13 @@ ANALYZE TABLE t1 PERSISTENT FOR COLUMNS(b) INDEXES(idx1, idx4);
 Table	Op	Msg_type	Msg_text
 test.t1	analyze	status	OK
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -576,12 +617,12 @@ t1	CREATE TABLE `t1` (
   KEY `idx4` (`e`,`x`,`d`)
 ) ENGINE=MyISAM DEFAULT CHARSET=latin1
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -605,12 +646,12 @@ t1	CREATE TABLE `t1` (
   KEY `idx4` (`e`,`b`,`d`)
 ) ENGINE=MyISAM DEFAULT CHARSET=latin1
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -624,13 +665,13 @@ LOAD DATA INFILE 'MYSQLTEST_VARDIR/tmp/save_index_stats'
   INTO TABLE mysql.index_stats
 FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"' LINES TERMINATED BY '\n';
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -658,12 +699,12 @@ t1	CREATE TABLE `t1` (
   KEY `idx4` (`e`,`d`)
 ) ENGINE=MyISAM DEFAULT CHARSET=latin1
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -721,12 +762,12 @@ t1	CREATE TABLE `t1` (
   KEY `idx4` (`e`,`b`,`d`)
 ) ENGINE=MyISAM DEFAULT CHARSET=latin1
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -735,13 +776,13 @@ ANALYZE TABLE t1 PERSISTENT FOR COLUMNS(b) INDEXES(idx1, idx2, idx4);
 Table	Op	Msg_type	Msg_text
 test.t1	analyze	status	OK
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	NULL	NULL	1.0000	NULL	NULL
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	NULL	NULL	1.0000	NULL	NULL	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -758,13 +799,13 @@ ANALYZE TABLE t1 PERSISTENT FOR COLUMNS(b) INDEXES(idx1, idx2, idx4);
 Table	Op	Msg_type	Msg_text
 test.t1	analyze	status	OK
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -790,12 +831,12 @@ t1	CREATE TABLE `t1` (
   KEY `idx3` (`d`)
 ) ENGINE=MyISAM DEFAULT CHARSET=latin1
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -820,12 +861,12 @@ t1	CREATE TABLE `t1` (
   KEY `idx4` (`e`,`b`,`d`)
 ) ENGINE=MyISAM DEFAULT CHARSET=latin1
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -834,13 +875,13 @@ ANALYZE TABLE t1 PERSISTENT FOR COLUMNS(b) INDEXES(idx1, idx2, idx4);
 Table	Op	Msg_type	Msg_text
 test.t1	analyze	status	OK
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -862,7 +903,7 @@ SELECT * FROM mysql.table_stats;
 db_name	table_name	cardinality
 test	t1	40
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 ANALYZE TABLE t1 PERSISTENT FOR COLUMNS(c,e,b) INDEXES(idx2,idx4);
@@ -872,10 +913,10 @@ SELECT * FROM mysql.table_stats;
 db_name	table_name	cardinality
 test	t1	40
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	idx2	1	7.0000
@@ -912,13 +953,13 @@ SELECT * FROM mysql.table_stats;
 db_name	table_name	cardinality
 test	t1	40
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -942,19 +983,19 @@ db_name	table_name	cardinality
 test	t1	40
 test	t2	40
 SELECT * FROM mysql.column_stats ORDER BY column_name;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t2	a	0	49	0.0000	4.0000	1.0000
-test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t2	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t2	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t2	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t2	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
-test	t2	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t2	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t2	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t2	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t2	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t2	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
+test	t2	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats ORDER BY index_name, prefix_arity, table_name;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -986,13 +1027,13 @@ SELECT * FROM mysql.table_stats;
 db_name	table_name	cardinality
 test	t2	40
 SELECT * FROM mysql.column_stats ORDER BY column_name;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t2	a	0	49	0.0000	4.0000	1.0000
-test	t2	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000
-test	t2	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t2	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t2	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t2	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t2	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t2	b	vvvvvvvvvvvvv	zzzzzzzzzzzzzzzzzz	0.2000	17.1250	6.4000	0	NULL	NULL
+test	t2	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t2	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t2	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t2	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats ORDER BY index_name, prefix_arity, table_name;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t2	PRIMARY	1	1.0000
@@ -1122,12 +1163,12 @@ MODIFY COLUMN b text,
 ADD INDEX idx1 (b(4), e), 
 ADD INDEX idx4 (e, b(4), d);
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t2	a	0	49	0.0000	4.0000	1.0000
-test	t2	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t2	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t2	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t2	f	1	5	0.2000	1.0000	6.4000
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t2	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t2	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t2	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t2	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t2	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t2	idx3	1	8.5000
@@ -1143,18 +1184,18 @@ ANALYZE TABLE t1;
 Table	Op	Msg_type	Msg_text
 test.t1	analyze	status	OK
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t2	a	0	49	0.0000	4.0000	1.0000
-test	t2	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t2	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t2	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t2	f	1	5	0.2000	1.0000	6.4000
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
-test	t1	b	NULL	NULL	0.2000	17.1250	NULL
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t2	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t2	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t2	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t2	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t2	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
+test	t1	b	NULL	NULL	0.2000	17.1250	NULL	NULL	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t2	idx3	1	8.5000
@@ -1187,7 +1228,7 @@ mysql.column_stats	analyze	status	OK
 SELECT * FROM mysql.table_stats;
 db_name	table_name	cardinality
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 set use_stat_tables='never';
@@ -1198,13 +1239,13 @@ SELECT * FROM mysql.table_stats;
 db_name	table_name	cardinality
 test	t1	40
 SELECT * FROM mysql.column_stats;
-db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency
-test	t1	a	0	49	0.0000	4.0000	1.0000
-test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000
-test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000
-test	t1	e	0.01	0.112	0.2250	8.0000	6.2000
-test	t1	f	1	5	0.2000	1.0000	6.4000
-test	t1	b	NULL	NULL	0.2000	17.1250	NULL
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_length	avg_frequency	hist_size	hist_type	histogram
+test	t1	a	0	49	0.0000	4.0000	1.0000	0	NULL	NULL
+test	t1	c	aaaa	dddddddd	0.1250	6.6571	7.0000	0	NULL	NULL
+test	t1	d	1989-03-12	1999-07-23	0.1500	3.0000	8.5000	0	NULL	NULL
+test	t1	e	0.01	0.112	0.2250	8.0000	6.2000	0	NULL	NULL
+test	t1	f	1	5	0.2000	1.0000	6.4000	0	NULL	NULL
+test	t1	b	NULL	NULL	0.2000	17.1250	NULL	NULL	NULL	NULL
 SELECT * FROM mysql.index_stats;
 db_name	table_name	index_name	prefix_arity	avg_frequency
 test	t1	PRIMARY	1	1.0000
@@ -1426,4 +1467,87 @@ UPPER(db_name)	UPPER(table_name)	index_name	prefix_arity	avg_frequency
 DELETE FROM mysql.table_stats;
 DELETE FROM mysql.column_stats;
 DELETE FROM mysql.index_stats;
+#
+# Bug mdev-4357: empty string as a value of the HIST_SIZE column 
+#                from mysql.column_stats
+#
+create table t1 (a int);
+insert into t1 values (1),(2),(3);
+set histogram_size=10;
+analyze table t1  persistent for all;
+Table	Op	Msg_type	Msg_text
+test.t1	analyze	status	OK
+select db_name, table_name, column_name,
+min_value, max_value,
+nulls_ratio, avg_frequency,
+hist_size, hist_type, HEX(histogram)
+FROM mysql.column_stats;
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_frequency	hist_size	hist_type	HEX(histogram)
+test	t1	a	1	3	0.0000	1.0000	10	SINGLE_PREC_HB	0000007F7F7F7FFFFFFF
+set histogram_size=default;
+drop table t1;
+#
+# Bug mdev-4359: wrong setting of the HIST_SIZE column 
+# (see also mdev-4357)         from mysql.column_stats
+#
+create table t1 ( a int);
+insert into t1 values (1),(2),(3),(4),(5);
+set histogram_size=10;
+set histogram_type='double_prec_hb';
+show variables like 'histogram%';
+Variable_name	Value
+histogram_size	10
+histogram_type	DOUBLE_PREC_HB
+analyze table t1 persistent for all;
+Table	Op	Msg_type	Msg_text
+test.t1	analyze	status	OK
+select db_name, table_name, column_name,
+min_value, max_value,
+nulls_ratio, avg_frequency,
+hist_size, hist_type, HEX(histogram)
+FROM mysql.column_stats;
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_frequency	hist_size	hist_type	HEX(histogram)
+test	t1	a	1	5	0.0000	1.0000	10	DOUBLE_PREC_HB	0000FF3FFF7FFFBFFFFF
+set histogram_size=default;
+set histogram_type=default;
+drop table t1;
+#
+# Bug mdev-4369: histogram for a column with many distinct values 
+#
+CREATE TABLE t1 (id int);
+CREATE TABLE t2 (id int);
+INSERT INTO t1 (id) VALUES (1), (1), (1),(1);
+INSERT INTO t1 (id) SELECT id FROM t1;
+INSERT INTO t1 SELECT id+1 FROM t1;
+INSERT INTO t1 SELECT id+2 FROM t1;
+INSERT INTO t1 SELECT id+4 FROM t1;
+INSERT INTO t1 SELECT id+8 FROM t1;
+INSERT INTO t1 SELECT id+16 FROM t1;
+INSERT INTO t1 SELECT id+32 FROM t1;
+INSERT INTO t1 SELECT id+64 FROM t1;
+INSERT INTO t1 SELECT id+128 FROM t1;
+INSERT INTO t1 SELECT id+256 FROM t1;
+INSERT INTO t1 SELECT id+512 FROM t1;
+INSERT INTO t2 SELECT id FROM t1 ORDER BY id*rand();
+SELECT COUNT(*) FROM t2;
+COUNT(*)
+8192
+SELECT COUNT(DISTINCT id) FROM t2;
+COUNT(DISTINCT id)
+1024
+set @@tmp_table_size=1024*16;
+set @@max_heap_table_size=1024*16;
+set histogram_size=63;
+analyze table t2 persistent for all;
+Table	Op	Msg_type	Msg_text
+test.t2	analyze	status	OK
+select db_name, table_name, column_name,
+min_value, max_value,
+nulls_ratio, avg_frequency,
+hist_size, hist_type, HEX(histogram)
+FROM mysql.column_stats;
+db_name	table_name	column_name	min_value	max_value	nulls_ratio	avg_frequency	hist_size	hist_type	HEX(histogram)
+test	t2	id	1	1024	0.0000	8.0000	63	SINGLE_PREC_HB	03070B0F13171B1F23272B2F33373B3F43474B4F53575B5F63676B6F73777B7F83878B8F93979B9FA3A7ABAFB3B7BBBFC3C7CBCFD3D7DBDFE3E7EBEFF3F7FB
+set histogram_size=default;
+drop table t1, t2;
 set use_stat_tables=@save_use_stat_tables;

mysql-test/r/system_mysql_db.result

@@ -287,6 +287,9 @@ column_stats	CREATE TABLE `column_stats` (
   `nulls_ratio` decimal(12,4) DEFAULT NULL,
   `avg_length` decimal(12,4) DEFAULT NULL,
   `avg_frequency` decimal(12,4) DEFAULT NULL,
+  `hist_size` tinyint(3) unsigned DEFAULT NULL,
+  `hist_type` enum('SINGLE_PREC_HB','DOUBLE_PREC_HB') COLLATE utf8_bin DEFAULT NULL,
+  `histogram` varbinary(255) DEFAULT NULL,
   PRIMARY KEY (`db_name`,`table_name`,`column_name`)
 ) ENGINE=MyISAM DEFAULT CHARSET=utf8 COLLATE=utf8_bin COMMENT='Statistics on Columns'
 show create table index_stats;

mysql-test/suite/funcs_1/r/is_columns_mysql.result

@@ -13,6 +13,9 @@ def	mysql	column_stats	avg_frequency	8	NULL	YES	decimal	NULL	NULL	12	4	NULL	NULL
 def	mysql	column_stats	avg_length	7	NULL	YES	decimal	NULL	NULL	12	4	NULL	NULL	NULL	decimal(12,4)			select,insert,update,references	
 def	mysql	column_stats	column_name	3	NULL	NO	varchar	64	192	NULL	NULL	NULL	utf8	utf8_bin	varchar(64)	PRI		select,insert,update,references	
 def	mysql	column_stats	db_name	1	NULL	NO	varchar	64	192	NULL	NULL	NULL	utf8	utf8_bin	varchar(64)	PRI		select,insert,update,references	
+def	mysql	column_stats	histogram	11	NULL	YES	varbinary	255	255	NULL	NULL	NULL	NULL	NULL	varbinary(255)			select,insert,update,references	
+def	mysql	column_stats	hist_size	9	NULL	YES	tinyint	NULL	NULL	3	0	NULL	NULL	NULL	tinyint(3) unsigned			select,insert,update,references	
+def	mysql	column_stats	hist_type	10	NULL	YES	enum	14	42	NULL	NULL	NULL	utf8	utf8_bin	enum('SINGLE_PREC_HB','DOUBLE_PREC_HB')			select,insert,update,references	
 def	mysql	column_stats	max_value	5	NULL	YES	varchar	255	765	NULL	NULL	NULL	utf8	utf8_bin	varchar(255)			select,insert,update,references	
 def	mysql	column_stats	min_value	4	NULL	YES	varchar	255	765	NULL	NULL	NULL	utf8	utf8_bin	varchar(255)			select,insert,update,references	
 def	mysql	column_stats	nulls_ratio	6	NULL	YES	decimal	NULL	NULL	12	4	NULL	NULL	NULL	decimal(12,4)			select,insert,update,references	
@@ -260,6 +263,7 @@ ORDER BY CHARACTER_SET_NAME, COLLATION_NAME, COL_CML;
 COL_CML	DATA_TYPE	CHARACTER_SET_NAME	COLLATION_NAME
 1.0000	blob	NULL	NULL
 1.0000	longblob	NULL	NULL
+1.0000	varbinary	NULL	NULL
 1.0000	char	latin1	latin1_bin
 1.0000	char	latin1	latin1_swedish_ci
 1.0000	varchar	latin1	latin1_swedish_ci
@@ -332,6 +336,9 @@ NULL	mysql	columns_priv	Timestamp	timestamp	NULL	NULL	NULL	NULL	timestamp
 NULL	mysql	column_stats	nulls_ratio	decimal	NULL	NULL	NULL	NULL	decimal(12,4)
 NULL	mysql	column_stats	avg_length	decimal	NULL	NULL	NULL	NULL	decimal(12,4)
 NULL	mysql	column_stats	avg_frequency	decimal	NULL	NULL	NULL	NULL	decimal(12,4)
+NULL	mysql	column_stats	hist_size	tinyint	NULL	NULL	NULL	NULL	tinyint(3) unsigned
+3.0000	mysql	column_stats	hist_type	enum	14	42	utf8	utf8_bin	enum('SINGLE_PREC_HB','DOUBLE_PREC_HB')
+1.0000	mysql	column_stats	histogram	varbinary	255	255	NULL	NULL	varbinary(255)
 3.0000	mysql	db	Host	char	60	180	utf8	utf8_bin	char(60)
 3.0000	mysql	db	Db	char	64	192	utf8	utf8_bin	char(64)
 3.0000	mysql	db	User	char	16	48	utf8	utf8_bin	char(16)

mysql-test/suite/funcs_1/r/is_columns_mysql_embedded.result

@@ -13,6 +13,9 @@ def	mysql	column_stats	avg_frequency	8	NULL	YES	decimal	NULL	NULL	12	4	NULL	NULL
 def	mysql	column_stats	avg_length	7	NULL	YES	decimal	NULL	NULL	12	4	NULL	NULL	NULL	decimal(12,4)				
 def	mysql	column_stats	column_name	3	NULL	NO	varchar	64	192	NULL	NULL	NULL	utf8	utf8_bin	varchar(64)	PRI			
 def	mysql	column_stats	db_name	1	NULL	NO	varchar	64	192	NULL	NULL	NULL	utf8	utf8_bin	varchar(64)	PRI			
+def	mysql	column_stats	histogram	11	NULL	YES	varbinary	255	255	NULL	NULL	NULL	NULL	NULL	varbinary(255)				
+def	mysql	column_stats	hist_size	9	NULL	YES	tinyint	NULL	NULL	3	0	NULL	NULL	NULL	tinyint(3) unsigned				
+def	mysql	column_stats	hist_type	10	NULL	YES	enum	14	42	NULL	NULL	NULL	utf8	utf8_bin	enum('SINGLE_PREC_HB','DOUBLE_PREC_HB')				
 def	mysql	column_stats	max_value	5	NULL	YES	varchar	255	765	NULL	NULL	NULL	utf8	utf8_bin	varchar(255)				
 def	mysql	column_stats	min_value	4	NULL	YES	varchar	255	765	NULL	NULL	NULL	utf8	utf8_bin	varchar(255)				
 def	mysql	column_stats	nulls_ratio	6	NULL	YES	decimal	NULL	NULL	12	4	NULL	NULL	NULL	decimal(12,4)				
@@ -260,6 +263,7 @@ ORDER BY CHARACTER_SET_NAME, COLLATION_NAME, COL_CML;
 COL_CML	DATA_TYPE	CHARACTER_SET_NAME	COLLATION_NAME
 1.0000	blob	NULL	NULL
 1.0000	longblob	NULL	NULL
+1.0000	varbinary	NULL	NULL
 1.0000	char	latin1	latin1_bin
 1.0000	char	latin1	latin1_swedish_ci
 1.0000	varchar	latin1	latin1_swedish_ci
@@ -332,6 +336,9 @@ NULL	mysql	columns_priv	Timestamp	timestamp	NULL	NULL	NULL	NULL	timestamp
 NULL	mysql	column_stats	nulls_ratio	decimal	NULL	NULL	NULL	NULL	decimal(12,4)
 NULL	mysql	column_stats	avg_length	decimal	NULL	NULL	NULL	NULL	decimal(12,4)
 NULL	mysql	column_stats	avg_frequency	decimal	NULL	NULL	NULL	NULL	decimal(12,4)
+NULL	mysql	column_stats	hist_size	tinyint	NULL	NULL	NULL	NULL	tinyint(3) unsigned
+3.0000	mysql	column_stats	hist_type	enum	14	42	utf8	utf8_bin	enum('SINGLE_PREC_HB','DOUBLE_PREC_HB')
+1.0000	mysql	column_stats	histogram	varbinary	255	255	NULL	NULL	varbinary(255)
 3.0000	mysql	db	Host	char	60	180	utf8	utf8_bin	char(60)
 3.0000	mysql	db	Db	char	64	192	utf8	utf8_bin	char(64)
 3.0000	mysql	db	User	char	16	48	utf8	utf8_bin	char(16)

mysql-test/suite/sys_vars/r/histogram_size_basic.result

@@ -0,0 +1,136 @@
+SET @start_global_value = @@global.histogram_size;
+SELECT @start_global_value;
+@start_global_value
+0
+SET @start_session_value = @@session.histogram_size;
+SELECT @start_session_value;
+@start_session_value
+0
+'#--------------------FN_DYNVARS_053_01-------------------------#'
+SET @@global.histogram_size = DEFAULT;
+SELECT @@global.histogram_size;
+@@global.histogram_size
+0
+SET @@session.histogram_size = DEFAULT;
+SELECT @@session.histogram_size;
+@@session.histogram_size
+0
+'#--------------------FN_DYNVARS_053_03-------------------------#'
+SET @@global.histogram_size = 1;
+SELECT @@global.histogram_size;
+@@global.histogram_size
+1
+SET @@global.histogram_size = 31;
+SELECT @@global.histogram_size;
+@@global.histogram_size
+31
+SET @@global.histogram_size = 255;
+SELECT @@global.histogram_size;
+@@global.histogram_size
+255
+'#--------------------FN_DYNVARS_053_04-------------------------#'
+SET @@session.histogram_size = 1;
+SELECT @@session.histogram_size;
+@@session.histogram_size
+1
+SET @@session.histogram_size = 31;
+SELECT @@session.histogram_size;
+@@session.histogram_size
+31
+SET @@session.histogram_size = 255;
+SELECT @@session.histogram_size;
+@@session.histogram_size
+255
+'#------------------FN_DYNVARS_053_05-----------------------#'
+SET @@global.histogram_size = -1;
+Warnings:
+Warning	1292	Truncated incorrect histogram_size value: '-1'
+SELECT @@global.histogram_size;
+@@global.histogram_size
+0
+SET @@global.histogram_size = 256;
+Warnings:
+Warning	1292	Truncated incorrect histogram_size value: '256'
+SELECT @@global.histogram_size;
+@@global.histogram_size
+255
+SET @@global.histogram_size = 1024;
+Warnings:
+Warning	1292	Truncated incorrect histogram_size value: '1024'
+SELECT @@global.histogram_size;
+@@global.histogram_size
+255
+SET @@global.histogram_size = 4.5;
+ERROR 42000: Incorrect argument type to variable 'histogram_size'
+SELECT @@global.histogram_size;
+@@global.histogram_size
+255
+SET @@global.histogram_size = test;
+ERROR 42000: Incorrect argument type to variable 'histogram_size'
+SELECT @@global.histogram_size;
+@@global.histogram_size
+255
+SET @@session.histogram_size = -1;
+Warnings:
+Warning	1292	Truncated incorrect histogram_size value: '-1'
+SELECT @@session.histogram_size;
+@@session.histogram_size
+0
+SET @@session.histogram_size = 256;
+Warnings:
+Warning	1292	Truncated incorrect histogram_size value: '256'
+SELECT @@session.histogram_size;
+@@session.histogram_size
+255
+SET @@session.histogram_size = 1024;
+Warnings:
+Warning	1292	Truncated incorrect histogram_size value: '1024'
+SELECT @@session.histogram_size;
+@@session.histogram_size
+255
+SET @@session.histogram_size = 4.5;
+ERROR 42000: Incorrect argument type to variable 'histogram_size'
+SELECT @@session.histogram_size;
+@@session.histogram_size
+255
+SET @@session.histogram_size = test;
+ERROR 42000: Incorrect argument type to variable 'histogram_size'
+SELECT @@session.histogram_size;
+@@session.histogram_size
+255
+'#------------------FN_DYNVARS_053_06-----------------------#'
+SELECT @@global.histogram_size = VARIABLE_VALUE 
+FROM INFORMATION_SCHEMA.GLOBAL_VARIABLES 
+WHERE VARIABLE_NAME='histogram_size';
+@@global.histogram_size = VARIABLE_VALUE
+1
+'#------------------FN_DYNVARS_053_07-----------------------#'
+SELECT @@session.histogram_size = VARIABLE_VALUE 
+FROM INFORMATION_SCHEMA.SESSION_VARIABLES
+WHERE VARIABLE_NAME='histogram_size';
+@@session.histogram_size = VARIABLE_VALUE
+1
+'#------------------FN_DYNVARS_053_08-----------------------#'
+SET @@global.histogram_size = TRUE;
+SET @@global.histogram_size = FALSE;
+'#---------------------FN_DYNVARS_001_09----------------------#'
+SET @@global.histogram_size = 10;
+SELECT @@histogram_size = @@global.histogram_size;
+@@histogram_size = @@global.histogram_size
+0
+'#---------------------FN_DYNVARS_001_10----------------------#'
+SET @@histogram_size = 100;
+SELECT @@histogram_size = @@local.histogram_size;
+@@histogram_size = @@local.histogram_size
+1
+SELECT @@local.histogram_size = @@session.histogram_size;
+@@local.histogram_size = @@session.histogram_size
+1
+SET @@global.histogram_size = @start_global_value;
+SELECT @@global.histogram_size;
+@@global.histogram_size
+0
+SET @@session.histogram_size = @start_session_value;
+SELECT @@session.histogram_size;
+@@session.histogram_size
+0

mysql-test/suite/sys_vars/r/histogram_type_basic.result

@@ -0,0 +1,79 @@
+SET @start_global_value = @@global.histogram_type;
+SELECT @start_global_value;
+@start_global_value
+SINGLE_PREC_HB
+SET @start_session_value = @@session.histogram_type;
+SELECT @start_session_value;
+@start_session_value
+SINGLE_PREC_HB
+SET @@global.histogram_type = 1;
+SET @@global.histogram_type = DEFAULT;
+SELECT @@global.histogram_type;
+@@global.histogram_type
+SINGLE_PREC_HB
+SET @@global.histogram_type = 0;
+SELECT @@global.histogram_type;
+@@global.histogram_type
+SINGLE_PREC_HB
+SET @@global.histogram_type = 1;
+SELECT @@global.histogram_type;
+@@global.histogram_type
+DOUBLE_PREC_HB
+SET @@global.histogram_type = SINGLE_PREC_HB;
+SELECT @@global.histogram_type;
+@@global.histogram_type
+SINGLE_PREC_HB
+SET @@global.histogram_type = DOUBLE_PREC_HB;
+SELECT @@global.histogram_type;
+@@global.histogram_type
+DOUBLE_PREC_HB
+SET @@session.histogram_type = 0;
+SELECT @@session.histogram_type;
+@@session.histogram_type
+SINGLE_PREC_HB
+SET @@session.histogram_type = 1;
+SELECT @@session.histogram_type;
+@@session.histogram_type
+DOUBLE_PREC_HB
+SET @@session.histogram_type = SINGLE_PREC_HB;
+SELECT @@session.histogram_type;
+@@session.histogram_type
+SINGLE_PREC_HB
+SET @@session.histogram_type = DOUBLE_PREC_HB;
+SELECT @@session.histogram_type;
+@@session.histogram_type
+DOUBLE_PREC_HB
+set sql_mode=TRADITIONAL;
+SET @@global.histogram_type = 10;
+ERROR 42000: Variable 'histogram_type' can't be set to the value of '10'
+SET @@global.histogram_type = -1024;
+ERROR 42000: Variable 'histogram_type' can't be set to the value of '-1024'
+SET @@global.histogram_type = 2.4;
+ERROR 42000: Incorrect argument type to variable 'histogram_type'
+SET @@global.histogram_type = OFF;
+ERROR 42000: Variable 'histogram_type' can't be set to the value of 'OFF'
+SET @@session.histogram_type = 10;
+ERROR 42000: Variable 'histogram_type' can't be set to the value of '10'
+SET @@session.histogram_type = -2;
+ERROR 42000: Variable 'histogram_type' can't be set to the value of '-2'
+SET @@session.histogram_type = 1.2;
+ERROR 42000: Incorrect argument type to variable 'histogram_type'
+SET @@session.histogram_type = ON;
+ERROR 42000: Variable 'histogram_type' can't be set to the value of 'ON'
+SELECT * FROM INFORMATION_SCHEMA.GLOBAL_VARIABLES 
+WHERE VARIABLE_NAME='histogram_type';
+VARIABLE_NAME	VARIABLE_VALUE
+HISTOGRAM_TYPE	DOUBLE_PREC_HB
+SELECT * FROM INFORMATION_SCHEMA.SESSION_VARIABLES 
+WHERE VARIABLE_NAME='histogram_type';
+VARIABLE_NAME	VARIABLE_VALUE
+HISTOGRAM_TYPE	DOUBLE_PREC_HB
+SET @@global.histogram_type = @start_global_value;
+SELECT @@global.histogram_type;
+@@global.histogram_type
+SINGLE_PREC_HB
+SET @@session.histogram_type = @start_session_value;
+SELECT @@session.histogram_type;
+@@session.histogram_type
+SINGLE_PREC_HB
+set sql_mode='';

mysql-test/suite/sys_vars/r/optimizer_use_condition_selectivity_basic.result

@@ -0,0 +1,109 @@
+SET @start_global_value = @@global.optimizer_use_condition_selectivity;
+SELECT @start_global_value;
+@start_global_value
+1
+SET @start_session_value = @@session.optimizer_use_condition_selectivity;
+SELECT @start_session_value;
+@start_session_value
+1
+'#--------------------FN_DYNVARS_115_01-------------------------#'
+SET @@global.optimizer_use_condition_selectivity = DEFAULT;
+SELECT @@global.optimizer_use_condition_selectivity;
+@@global.optimizer_use_condition_selectivity
+1
+SET @@session.optimizer_use_condition_selectivity = DEFAULT;
+SELECT @@session.optimizer_use_condition_selectivity;
+@@session.optimizer_use_condition_selectivity
+1
+'#--------------------FN_DYNVARS_115_02-------------------------#'
+SET @@global.optimizer_use_condition_selectivity = DEFAULT;
+SELECT @@global.optimizer_use_condition_selectivity = 1;
+@@global.optimizer_use_condition_selectivity = 1
+1
+SET @@session.optimizer_use_condition_selectivity = DEFAULT;
+SELECT @@session.optimizer_use_condition_selectivity = 1;
+@@session.optimizer_use_condition_selectivity = 1
+1
+'#--------------------FN_DYNVARS_115_03-------------------------#'
+SELECT @@global.optimizer_use_condition_selectivity;
+@@global.optimizer_use_condition_selectivity
+1
+SET @@global.optimizer_use_condition_selectivity = 1;
+SELECT @@global.optimizer_use_condition_selectivity;
+@@global.optimizer_use_condition_selectivity
+1
+SET @@global.optimizer_use_condition_selectivity = 2;
+SELECT @@global.optimizer_use_condition_selectivity;
+@@global.optimizer_use_condition_selectivity
+2
+SET @@global.optimizer_use_condition_selectivity = 3;
+SELECT @@global.optimizer_use_condition_selectivity;
+@@global.optimizer_use_condition_selectivity
+3
+SET @@global.optimizer_use_condition_selectivity = 4;
+SELECT @@global.optimizer_use_condition_selectivity;
+@@global.optimizer_use_condition_selectivity
+4
+'#--------------------FN_DYNVARS_115_04-------------------------#'
+SELECT @@session.optimizer_use_condition_selectivity;
+@@session.optimizer_use_condition_selectivity
+1
+SET @@session.optimizer_use_condition_selectivity = 1;
+SELECT @@session.optimizer_use_condition_selectivity;
+@@session.optimizer_use_condition_selectivity
+1
+SET @@session.optimizer_use_condition_selectivity = 2;
+SELECT @@session.optimizer_use_condition_selectivity;
+@@session.optimizer_use_condition_selectivity
+2
+SET @@session.optimizer_use_condition_selectivity = 3;
+SELECT @@session.optimizer_use_condition_selectivity;
+@@session.optimizer_use_condition_selectivity
+3
+SET @@session.optimizer_use_condition_selectivity = 4;
+SELECT @@session.optimizer_use_condition_selectivity;
+@@session.optimizer_use_condition_selectivity
+4
+'#------------------FN_DYNVARS_115_05-----------------------#'
+SET @@global.optimizer_use_condition_selectivity = ON;
+ERROR 42000: Incorrect argument type to variable 'optimizer_use_condition_selectivity'
+SET @@global.optimizer_use_condition_selectivity = OFF;
+ERROR 42000: Incorrect argument type to variable 'optimizer_use_condition_selectivity'
+SET @@session.optimizer_use_condition_selectivity = 65530.34;
+ERROR 42000: Incorrect argument type to variable 'optimizer_use_condition_selectivity'
+SET @@session.optimizer_use_condition_selectivity = test;
+ERROR 42000: Incorrect argument type to variable 'optimizer_use_condition_selectivity'
+'#------------------FN_DYNVARS_115_06-----------------------#'
+SELECT @@global.optimizer_use_condition_selectivity = VARIABLE_VALUE 
+FROM INFORMATION_SCHEMA.GLOBAL_VARIABLES 
+WHERE VARIABLE_NAME='optimizer_use_condition_selectivity';
+@@global.optimizer_use_condition_selectivity = VARIABLE_VALUE
+1
+'#------------------FN_DYNVARS_115_07-----------------------#'
+SELECT @@session.optimizer_use_condition_selectivity = VARIABLE_VALUE
+FROM INFORMATION_SCHEMA.SESSION_VARIABLES 
+WHERE VARIABLE_NAME='optimizer_use_condition_selectivity';
+@@session.optimizer_use_condition_selectivity = VARIABLE_VALUE
+1
+'#---------------------FN_DYNVARS_115_08----------------------#'
+SET @@optimizer_use_condition_selectivity = 1;
+SET @@global.optimizer_use_condition_selectivity = 3;
+SELECT @@optimizer_use_condition_selectivity = @@global.optimizer_use_condition_selectivity;
+@@optimizer_use_condition_selectivity = @@global.optimizer_use_condition_selectivity
+0
+'#---------------------FN_DYNVARS_115_09----------------------#'
+SET @@optimizer_use_condition_selectivity = 2;
+SELECT @@optimizer_use_condition_selectivity = @@local.optimizer_use_condition_selectivity;
+@@optimizer_use_condition_selectivity = @@local.optimizer_use_condition_selectivity
+1
+SELECT @@local.optimizer_use_condition_selectivity = @@session.optimizer_use_condition_selectivity;
+@@local.optimizer_use_condition_selectivity = @@session.optimizer_use_condition_selectivity
+1
+SET @@global.optimizer_use_condition_selectivity = @start_global_value;
+SELECT @@global.optimizer_use_condition_selectivity;
+@@global.optimizer_use_condition_selectivity
+1
+SET @@session.optimizer_use_condition_selectivity = @start_session_value;
+SELECT @@session.optimizer_use_condition_selectivity;
+@@session.optimizer_use_condition_selectivity
+1

mysql-test/suite/sys_vars/t/histogram_size_basic.test

@@ -0,0 +1,142 @@
+--source include/load_sysvars.inc
+
+##############################################################
+#           START OF histogram_size TESTS                    #
+##############################################################
+
+
+#############################################################
+#                 Save initial value                        #
+#############################################################
+
+SET @start_global_value = @@global.histogram_size;
+SELECT @start_global_value;
+SET @start_session_value = @@session.histogram_size;
+SELECT @start_session_value;
+
+--echo '#--------------------FN_DYNVARS_053_01-------------------------#'
+################################################################
+#     Display the DEFAULT value of histogram_size            #
+################################################################
+
+SET @@global.histogram_size = DEFAULT;
+SELECT @@global.histogram_size;
+
+SET @@session.histogram_size = DEFAULT;
+SELECT @@session.histogram_size;
+
+--echo '#--------------------FN_DYNVARS_053_03-------------------------#'
+########################################################################
+# Change the value of histogram_size to a valid value for GLOBAL Scope #
+########################################################################
+
+SET @@global.histogram_size = 1;
+SELECT @@global.histogram_size;
+SET @@global.histogram_size = 31;
+SELECT @@global.histogram_size;
+SET @@global.histogram_size = 255;
+SELECT @@global.histogram_size;
+
+--echo '#--------------------FN_DYNVARS_053_04-------------------------#'
+#########################################################################
+# Change the value of histogram_size to a valid value for SESSION Scope #
+#########################################################################
+ 
+SET @@session.histogram_size = 1;
+SELECT @@session.histogram_size;
+SET @@session.histogram_size = 31;
+SELECT @@session.histogram_size;
+SET @@session.histogram_size = 255;
+SELECT @@session.histogram_size;
+
+--echo '#------------------FN_DYNVARS_053_05-----------------------#'
+##########################################################
+# Change the value of histogram_size to an invalid value #
+###########################################################
+
+SET @@global.histogram_size = -1;
+SELECT @@global.histogram_size;
+SET @@global.histogram_size = 256;
+SELECT @@global.histogram_size;
+SET @@global.histogram_size = 1024;
+SELECT @@global.histogram_size;
+
+--Error ER_WRONG_TYPE_FOR_VAR
+SET @@global.histogram_size = 4.5;
+SELECT @@global.histogram_size;
+--Error ER_WRONG_TYPE_FOR_VAR
+SET @@global.histogram_size = test;
+SELECT @@global.histogram_size;
+
+SET @@session.histogram_size = -1;
+SELECT @@session.histogram_size;
+SET @@session.histogram_size = 256;
+SELECT @@session.histogram_size;
+SET @@session.histogram_size = 1024;
+SELECT @@session.histogram_size;
+
+--Error ER_WRONG_TYPE_FOR_VAR
+SET @@session.histogram_size = 4.5;
+SELECT @@session.histogram_size;
+--Error ER_WRONG_TYPE_FOR_VAR
+SET @@session.histogram_size = test;
+SELECT @@session.histogram_size;
+
+--echo '#------------------FN_DYNVARS_053_06-----------------------#'
+####################################################################
+#   Check if the value in GLOBAL Table matches value in variable   #
+####################################################################
+
+SELECT @@global.histogram_size = VARIABLE_VALUE 
+FROM INFORMATION_SCHEMA.GLOBAL_VARIABLES 
+WHERE VARIABLE_NAME='histogram_size';
+
+--echo '#------------------FN_DYNVARS_053_07-----------------------#'
+####################################################################
+#  Check if the value in SESSION Table matches value in variable   #
+####################################################################
+
+SELECT @@session.histogram_size = VARIABLE_VALUE 
+FROM INFORMATION_SCHEMA.SESSION_VARIABLES
+WHERE VARIABLE_NAME='histogram_size';
+
+
+--echo '#------------------FN_DYNVARS_053_08-----------------------#'
+####################################################################
+#     Check if TRUE and FALSE values can be used on variable       #
+####################################################################
+
+SET @@global.histogram_size = TRUE;
+SET @@global.histogram_size = FALSE;
+
+--echo '#---------------------FN_DYNVARS_001_09----------------------#'
+#################################################################################
+#  Check if accessing variable with and without GLOBAL point to same variable   #
+#################################################################################
+
+SET @@global.histogram_size = 10;
+SELECT @@histogram_size = @@global.histogram_size;
+
+--echo '#---------------------FN_DYNVARS_001_10----------------------#'
+########################################################################################################
+#    Check if accessing variable with SESSION,LOCAL and without SCOPE points to same session variable  #
+########################################################################################################
+
+SET @@histogram_size = 100;
+SELECT @@histogram_size = @@local.histogram_size;
+SELECT @@local.histogram_size = @@session.histogram_size;
+
+####################################
+#     Restore initial value        #
+####################################
+
+SET @@global.histogram_size = @start_global_value;
+SELECT @@global.histogram_size;
+SET @@session.histogram_size = @start_session_value;
+SELECT @@session.histogram_size;
+
+
+###################################################
+#           END OF histogram_size TESTS           #
+###################################################
+

mysql-test/suite/sys_vars/t/histogram_type_basic.test

@@ -0,0 +1,92 @@
+--source include/load_sysvars.inc
+
+#############################################################
+#                 Save initial value                        #
+#############################################################
+
+SET @start_global_value = @@global.histogram_type;
+SELECT @start_global_value;
+SET @start_session_value = @@session.histogram_type;
+SELECT @start_session_value;
+
+##############################################################
+#     Display the DEFAULT value of histogram_type            #
+##############################################################
+
+SET @@global.histogram_type = 1;
+SET @@global.histogram_type = DEFAULT;
+SELECT @@global.histogram_type;
+
+#################################################################################
+# Change the value of histogram_type to a valid value for GLOBAL Scope          #
+#################################################################################
+
+SET @@global.histogram_type = 0;
+SELECT @@global.histogram_type;
+SET @@global.histogram_type = 1;
+SELECT @@global.histogram_type;
+
+SET @@global.histogram_type = SINGLE_PREC_HB;
+SELECT @@global.histogram_type;
+SET @@global.histogram_type = DOUBLE_PREC_HB;
+SELECT @@global.histogram_type;
+
+###################################################################################
+# Change the value of histogram_type to a valid value for SESSION Scope           #
+###################################################################################
+
+SET @@session.histogram_type = 0;
+SELECT @@session.histogram_type;
+SET @@session.histogram_type = 1;
+SELECT @@session.histogram_type;
+
+SET @@session.histogram_type = SINGLE_PREC_HB;
+SELECT @@session.histogram_type;
+SET @@session.histogram_type = DOUBLE_PREC_HB;
+SELECT @@session.histogram_type;
+
+####################################################################
+# Change the value of histogram_type to an invalid value           #
+####################################################################
+
+set sql_mode=TRADITIONAL;
+--Error ER_WRONG_VALUE_FOR_VAR
+SET @@global.histogram_type = 10;
+--Error ER_WRONG_VALUE_FOR_VAR
+SET @@global.histogram_type = -1024;
+--Error ER_WRONG_TYPE_FOR_VAR
+SET @@global.histogram_type = 2.4;
+--Error ER_WRONG_VALUE_FOR_VAR
+SET @@global.histogram_type = OFF;
+--Error ER_WRONG_VALUE_FOR_VAR
+SET @@session.histogram_type = 10;
+--Error ER_WRONG_VALUE_FOR_VAR
+SET @@session.histogram_type = -2;
+--Error ER_WRONG_TYPE_FOR_VAR
+SET @@session.histogram_type = 1.2;
+--Error ER_WRONG_VALUE_FOR_VAR
+SET @@session.histogram_type = ON;
+
+###############################################################################
+#   Check if the value in GLOBAL & SESSION Tables matches value in variable   #
+###############################################################################
+
+SELECT * FROM INFORMATION_SCHEMA.GLOBAL_VARIABLES 
+WHERE VARIABLE_NAME='histogram_type';
+
+SELECT * FROM INFORMATION_SCHEMA.SESSION_VARIABLES 
+WHERE VARIABLE_NAME='histogram_type';
+
+####################################
+#     Restore initial value        #
+####################################
+
+SET @@global.histogram_type = @start_global_value;
+SELECT @@global.histogram_type;
+SET @@session.histogram_type = @start_session_value;
+SELECT @@session.histogram_type;
+set sql_mode='';
+
+#####################################################
+#         END OF histogram_type TESTS               #
+#####################################################

mysql-test/suite/sys_vars/t/optimizer_use_condition_selectivity_basic.test

@@ -0,0 +1,142 @@
+
+--source include/load_sysvars.inc
+
+#################################################################
+#      START OF optimizer_use_condition_selectivity TESTS       # 
+#################################################################
+
+
+#############################################################
+#                 Save initial value                        #
+#############################################################
+
+SET @start_global_value = @@global.optimizer_use_condition_selectivity;
+SELECT @start_global_value;
+SET @start_session_value = @@session.optimizer_use_condition_selectivity;
+SELECT @start_session_value;
+
+
+--echo '#--------------------FN_DYNVARS_115_01-------------------------#'
+#########################################################################
+#     Display the DEFAULT value of optimizer_use_condition_selectivity  #
+#########################################################################
+
+SET @@global.optimizer_use_condition_selectivity = DEFAULT;
+SELECT @@global.optimizer_use_condition_selectivity;
+
+SET @@session.optimizer_use_condition_selectivity = DEFAULT;
+SELECT @@session.optimizer_use_condition_selectivity;
+
+
+--echo '#--------------------FN_DYNVARS_115_02-------------------------#'
+#########################################################################
+#     Check the DEFAULT value of optimizer_use_condition_selectivity    #
+#########################################################################
+
+SET @@global.optimizer_use_condition_selectivity = DEFAULT;
+SELECT @@global.optimizer_use_condition_selectivity = 1;
+
+SET @@session.optimizer_use_condition_selectivity = DEFAULT;
+SELECT @@session.optimizer_use_condition_selectivity = 1;
+
+
+--echo '#--------------------FN_DYNVARS_115_03-------------------------#'
+#############################################################################################
+# Change the value of optimizer_use_condition_selectivity to a valid value for GLOBAL Scope #
+#############################################################################################
+
+SELECT @@global.optimizer_use_condition_selectivity;
+SET @@global.optimizer_use_condition_selectivity = 1;
+SELECT @@global.optimizer_use_condition_selectivity;
+SET @@global.optimizer_use_condition_selectivity = 2;
+SELECT @@global.optimizer_use_condition_selectivity;
+SET @@global.optimizer_use_condition_selectivity = 3;
+SELECT @@global.optimizer_use_condition_selectivity;
+SET @@global.optimizer_use_condition_selectivity = 4;
+SELECT @@global.optimizer_use_condition_selectivity;
+
+
+--echo '#--------------------FN_DYNVARS_115_04-------------------------#'
+#############################################################################################
+# Change the value of optimizer_use_condition_selectivity to a valid value for SESSION Scope#
+#############################################################################################
+ 
+SELECT @@session.optimizer_use_condition_selectivity;
+SET @@session.optimizer_use_condition_selectivity = 1;
+SELECT @@session.optimizer_use_condition_selectivity;
+SET @@session.optimizer_use_condition_selectivity = 2;
+SELECT @@session.optimizer_use_condition_selectivity;
+SET @@session.optimizer_use_condition_selectivity = 3;
+SELECT @@session.optimizer_use_condition_selectivity;
+SET @@session.optimizer_use_condition_selectivity = 4;
+SELECT @@session.optimizer_use_condition_selectivity;
+
+
+--echo '#------------------FN_DYNVARS_115_05-----------------------#'
+###############################################################################
+# Change the value of optimizer_use_condition_selectivity to an invalid value #
+##############################################################################
+
+--Error ER_WRONG_TYPE_FOR_VAR
+SET @@global.optimizer_use_condition_selectivity = ON;
+--Error ER_WRONG_TYPE_FOR_VAR
+SET @@global.optimizer_use_condition_selectivity = OFF;
+--Error ER_WRONG_TYPE_FOR_VAR
+SET @@session.optimizer_use_condition_selectivity = 65530.34;
+--Error ER_WRONG_TYPE_FOR_VAR
+SET @@session.optimizer_use_condition_selectivity = test;
+
+--echo '#------------------FN_DYNVARS_115_06-----------------------#'
+####################################################################
+#   Check if the value in GLOBAL Table matches value in variable   #
+####################################################################
+
+
+SELECT @@global.optimizer_use_condition_selectivity = VARIABLE_VALUE 
+FROM INFORMATION_SCHEMA.GLOBAL_VARIABLES 
+WHERE VARIABLE_NAME='optimizer_use_condition_selectivity';
+
+--echo '#------------------FN_DYNVARS_115_07-----------------------#'
+####################################################################
+#  Check if the value in SESSION Table matches value in variable   #
+####################################################################
+
+SELECT @@session.optimizer_use_condition_selectivity = VARIABLE_VALUE
+FROM INFORMATION_SCHEMA.SESSION_VARIABLES 
+WHERE VARIABLE_NAME='optimizer_use_condition_selectivity';
+
+
+--echo '#---------------------FN_DYNVARS_115_08----------------------#'
+###############################################################################
+#  Check if global and session variable are independent of each other         #
+###############################################################################
+
+SET @@optimizer_use_condition_selectivity = 1;
+SET @@global.optimizer_use_condition_selectivity = 3;
+SELECT @@optimizer_use_condition_selectivity = @@global.optimizer_use_condition_selectivity;
+
+
+--echo '#---------------------FN_DYNVARS_115_09----------------------#'
+###############################################################################
+#    Check if accessing variable with SESSION,LOCAL and without SCOPE points  #
+#    to same session variable                                                 #
+###############################################################################
+
+SET @@optimizer_use_condition_selectivity = 2;
+SELECT @@optimizer_use_condition_selectivity = @@local.optimizer_use_condition_selectivity;
+SELECT @@local.optimizer_use_condition_selectivity = @@session.optimizer_use_condition_selectivity;
+
+
+####################################
+#     Restore initial value        #
+####################################
+
+SET @@global.optimizer_use_condition_selectivity = @start_global_value;
+SELECT @@global.optimizer_use_condition_selectivity;
+SET @@session.optimizer_use_condition_selectivity = @start_session_value;
+SELECT @@session.optimizer_use_condition_selectivity;
+
+########################################################################
+#                 END OF optimizer_use_condition_selectivity TESTS     #
+########################################################################
+

mysql-test/t/selectivity.test

@@ -0,0 +1,623 @@
+--source include/have_stat_tables.inc
+
+select @@global.use_stat_tables;
+select @@session.use_stat_tables;
+
+set @save_use_stat_tables=@@use_stat_tables;
+
+set use_stat_tables='preferably';
+
+set @save_optimizer_use_condition_selectivity=@@optimizer_use_condition_selectivity;
+set @save_histogram_size=@@histogram_size;
+set @save_histogram_type=@@histogram_type;
+
+# check that statistics on nulls is used
+
+set optimizer_use_condition_selectivity=3;
+
+create table t1 (a int);
+insert into t1 values
+  (9), (3), (2), (NULL), (NULL), (2), (NULL), (1), (5), (NULL);
+
+analyze table t1;
+
+select * from mysql.column_stats;
+
+flush table t1;
+
+explain extended 
+select * from t1 where a is null;
+
+explain extended 
+select * from t1 where a is not null;
+
+drop table t1;
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+--disable_warnings
+DROP DATABASE IF EXISTS dbt3_s001;
+--enable_warnings
+
+CREATE DATABASE dbt3_s001;
+
+use dbt3_s001;
+
+--disable_query_log
+--disable_result_log
+--disable_warnings
+--source include/dbt3_s001.inc
+
+ANALYZE TABLE
+customer, lineitem, nation, orders, part, partsupp, region, supplier;
+FLUSH TABLE
+customer, lineitem, nation, orders, part, partsupp, region, supplier;
+--enable_warnings
+--enable_result_log
+--enable_query_log
+
+
+--echo === Q15 ===
+
+create view revenue0 (supplier_no, total_revenue) as
+  select l_suppkey, sum(l_extendedprice * (1 - l_discount))
+  from lineitem
+  where
+    l_shipdate >= '1995-08-01'
+    and l_shipdate < date_add('1995-08-01', interval 90 day)
+  group by l_suppkey;
+
+let $Q15=
+select s_suppkey, s_name, s_address, s_phone, total_revenue
+from supplier, revenue0
+where s_suppkey = supplier_no
+  and total_revenue = (select max(total_revenue) from revenue0)
+order by s_suppkey;
+
+set @save_optimizer_switch=@@optimizer_switch;
+set optimizer_switch='index_condition_pushdown=off';
+
+set optimizer_use_condition_selectivity=1;
+eval EXPLAIN EXTENDED $Q15;
+eval $Q15;
+
+set optimizer_use_condition_selectivity=3;
+eval EXPLAIN EXTENDED $Q15;
+eval $Q15;
+
+set optimizer_switch=@save_optimizer_switch;
+
+drop view revenue0;
+
+
+--echo === Q16 ===
+
+let $Q16=
+select p_brand, p_type, p_size, count(distinct ps_suppkey) as supplier_cnt
+from partsupp, part
+where p_partkey = ps_partkey
+  and p_brand <> 'Brand#11'
+  and p_type not like 'SMALL POLISHED%'
+  and p_size in (49, 37, 27, 5, 40, 6, 22, 8)
+  and ps_suppkey not in (select s_suppkey from supplier
+                           where s_comment like '%Customer%Complaints%')
+group by p_brand, p_type, p_size
+order by supplier_cnt desc, p_brand, p_type, p_size;
+
+set optimizer_use_condition_selectivity=1;
+eval EXPLAIN EXTENDED $Q16;
+eval $Q16;
+
+set optimizer_use_condition_selectivity=3;
+eval EXPLAIN EXTENDED $Q16;
+eval $Q16;
+
+set optimizer_use_condition_selectivity=4;
+eval EXPLAIN EXTENDED $Q16;
+eval $Q16;
+
+
+--echo === Q18 ===
+
+let $Q18=
+select 
+  c_name, c_custkey, o_orderkey, o_orderdate, o_totalprice, sum(l_quantity)
+from customer, orders, lineitem
+where
+  o_orderkey in (select l_orderkey from lineitem
+                 group by l_orderkey having sum(l_quantity) > 250)
+  and c_custkey = o_custkey
+  and o_orderkey = l_orderkey
+group by c_name, c_custkey, o_orderkey, o_orderdate, o_totalprice
+order by o_totalprice desc, o_orderdate;
+
+set optimizer_use_condition_selectivity=1;
+eval EXPLAIN EXTENDED $Q18;
+eval $Q18;
+
+set optimizer_use_condition_selectivity=3;
+eval EXPLAIN EXTENDED $Q18;
+eval $Q18;
+
+
+--echo === Q22 ===
+
+let $Q22=
+select cntrycode, count(*) as numcust, sum(c_acctbal) as totacctbal
+from  (
+    select substr(c_phone, 1, 2) as cntrycode, c_acctbal
+    from customer
+    where
+      substr(c_phone, 1, 2) in ('10', '20', '14', '19', '11', '28', '25')
+      and c_acctbal > (select avg(c_acctbal) from customer
+                        where c_acctbal > 0.00
+                          and substr(c_phone, 1, 2) in
+                              ('10', '20', '14', '19', '11', '28', '25'))
+      and not exists (select * from orders where o_custkey = c_custkey)
+  ) as vip
+group by cntrycode
+order by cntrycode;
+
+set optimizer_use_condition_selectivity=1;
+eval EXPLAIN EXTENDED $Q22;
+eval $Q22;
+
+set optimizer_use_condition_selectivity=3;
+eval EXPLAIN EXTENDED $Q22;
+eval $Q22;
+
+
+--echo === Q20 ===
+
+let $Q20=
+select sql_calc_found_rows
+       s_name, s_address
+from supplier, nation
+where s_suppkey in (select ps_suppkey from partsupp
+                    where ps_partkey in (select p_partkey from part
+                                         where p_name like 'g%')
+                          and ps_availqty >
+                              (select 0.5 * sum(l_quantity)
+                               from lineitem
+                               where l_partkey = ps_partkey
+                                     and l_suppkey = ps_suppkey
+                                     and l_shipdate >= date('1993-01-01')
+                                     and l_shipdate < date('1993-01-01') +
+                                         interval '1' year ))
+and s_nationkey = n_nationkey
+and n_name = 'UNITED STATES'
+order by s_name
+limit 10;
+
+set optimizer_use_condition_selectivity=1;
+eval EXPLAIN EXTENDED $Q20;
+eval $Q20;
+
+SELECT ((SELECT COUNT(*) FROM part WHERE p_name LIKE 'g%') /
+        (SELECT COUNT(*) FROM part)) AS sel;
+
+set optimizer_use_condition_selectivity=3;
+eval EXPLAIN EXTENDED $Q20;
+eval $Q20;
+
+set histogram_size=127;
+
+ANALYZE TABLE part PERSISTENT FOR COLUMNS(p_name) INDEXES();
+
+flush table part;
+
+set optimizer_use_condition_selectivity=4;
+eval EXPLAIN EXTENDED $Q20;
+eval $Q20; 
+
+set histogram_type='DOUBLE_PREC_HB';
+set histogram_size=126;
+
+ANALYZE TABLE part PERSISTENT FOR COLUMNS(p_name) INDEXES();
+
+flush table part;
+
+eval EXPLAIN EXTENDED $Q20;
+eval $Q20; 
+
+set histogram_type='SINGLE_PREC_HB';
+set histogram_size=24;
+
+ANALYZE TABLE nation PERSISTENT FOR COLUMNS(n_name) INDEXES();
+
+flush table nation;
+
+eval EXPLAIN EXTENDED $Q20;
+eval $Q20; 
+
+
+DROP DATABASE dbt3_s001;
+
+set histogram_type=@save_histogram_type;
+set histogram_size=@save_histogram_size;
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+use test;
+
+--echo #
+--echo # Bug mdev-4348: using view with use_condition_selectivity > 1 
+--echo #
+
+set @tmp_use_stat_tables=@@use_stat_tables; 
+set use_stat_tables='never';
+set optimizer_use_condition_selectivity=3;
+
+CREATE TABLE t1 (a int, b int);
+INSERT t1 VALUES (7,1), (0,7);
+CREATE ALGORITHM=MERGE VIEW v1 AS SELECT * FROM t1;
+
+CREATE TABLE t2 (c int, d int, index idx(d));
+INSERT INTO t2 VALUES 
+  (0,4), (8,6), (1,3), (8,5), (9,3), (2,2), (6,2),
+  (1,9), (6,3), (2,8), (4,1), (0,7), (4,8), (4,5);
+
+EXPLAIN EXTENDED
+SELECT * FROM v1 INNER JOIN t2 ON ( a = c AND b = d );
+
+SELECT * FROM v1 INNER JOIN t2 ON ( a = c AND b = d );
+
+DROP VIEW v1;
+DROP TABLE t1,t2;
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+set use_stat_tables=@tmp_use_stat_tables; 
+
+--echo #
+--echo # Bug mdev-4349: impossible range for non-indexed column 
+--echo #
+
+set optimizer_use_condition_selectivity=3;
+
+create table t1 (a int);
+insert into t1 values
+  (3), (7), (2), (5), (7), (1), (2), (2);
+
+set optimizer_use_condition_selectivity=1;
+explain extended
+select * from t1 where a < 1 and a > 7;
+select * from t1 where a < 1 and a > 7;
+
+set optimizer_use_condition_selectivity=3;
+explain extended
+select * from t1 where a < 1 and a > 7;
+select * from t1 where a < 1 and a > 7;
+
+drop table t1;
+
+create table t1 (a int);
+insert into t1 values (1);
+
+create table t2 (b int);
+insert into t2 values (2),(3);
+
+explain extended 
+select * from t1 where a in ( select b from t2 ) AND ( a > 3 );
+select * from t1 where a in ( select b from t2 ) AND ( a > 3 );
+
+drop table t1,t2;
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+--echo #
+--echo # Bug mdev-4350: erroneous negative selectivity 
+--echo #
+
+create table t1 (a int);
+insert into t1 values (1), (1);
+insert into t1 select * from t1; 
+insert into t1 select * from t1; 
+insert into t1 select * from t1; 
+insert into t1 select * from t1; 
+insert into t1 select * from t1; 
+insert into t1 select * from t1; 
+insert into t1 select * from t1; 
+insert into t1 select * from t1; 
+insert into t1 select * from t1; 
+insert into t1 values (0);
+select count(*) from t1;
+
+set use_stat_tables='preferably';
+set histogram_size=127;
+set histogram_type='SINGLE_PREC_HB';
+analyze table t1;
+flush table t1;
+
+set optimizer_use_condition_selectivity=4;
+explain extended select * from t1 where a=0;
+
+drop table t1;
+
+set histogram_size=@save_histogram_size;
+set histogram_type=@save_histogram_type;
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+--echo #
+--echo # Bug mdev-4367: 2-way join with an empty table 
+--echo #                when optimizer_use_condition_selectivity=3
+--echo #
+
+set optimizer_use_condition_selectivity=3;
+
+CREATE TABLE t1 (a varchar(1)) ENGINE=MyISAM;
+INSERT INTO t1 VALUES ('j'),('k');
+
+CREATE TABLE t2 (b varchar(1)) ENGINE=MyISAM;
+INSERT INTO t2 VALUES ('x'),('y');
+
+CREATE TABLE t3 (c varchar(1), KEY(c)) ENGINE=MyISAM;
+
+SELECT * FROM t1 STRAIGHT_JOIN (t2 JOIN t3 ON c = b AND b > 'z');
+
+DROP TABLE t1,t2,t3;
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+--echo #
+--echo # Bug mdev-4366: impossible condition on an indexed column discovered after
+--echo #                substitution of constant tables
+--echo #                with optimizer_use_condition_selectivity=3
+--echo #
+
+CREATE TABLE t1 (pk int PRIMARY KEY, a int);
+INSERT INTO t1 VALUES 
+  (1,4), (2,6), (3,3), (4,5);
+
+CREATE TABLE t2 (b int);
+INSERT INTO t2 VALUES (1), (7);
+
+set optimizer_use_condition_selectivity=1;
+EXPLAIN EXTENDED
+SELECT 1 FROM t1, t2 WHERE pk = 6 AND a = 2 AND b = 10;
+SELECT 1 FROM t1, t2 WHERE pk = 6 AND a = 2 AND b = 10;
+
+set optimizer_use_condition_selectivity=3;
+EXPLAIN EXTENDED
+SELECT 1 FROM t1, t2 WHERE pk = 6 AND a = 2 AND b = 10;
+SELECT 1 FROM t1, t2 WHERE pk = 6 AND a = 2 AND b = 10;
+
+DROP TABLE t1,t2;
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+--echo #
+--echo # Bug mdev-4370: Histograms have bean created, but the query is run after
+--echo #                FLUSH TABLES with optimizer_use_condition_selectivity=3
+--echo #
+
+set use_stat_tables=PREFERABLY; 
+set histogram_size=10;
+set histogram_type='SINGLE_PREC_HB';
+
+CREATE TABLE t1 (a int);
+INSERT INTO t1 VALUES (9), (1);
+ANALYZE TABLE t1;
+FLUSH TABLES;
+
+set optimizer_use_condition_selectivity=3;
+
+EXPLAIN EXTENDED
+SELECT * FROM t1 WHERE a > 3;
+SELECT * FROM t1 WHERE a > 3;
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+DROP TABLE t1;
+
+set histogram_size=@save_histogram_size;
+set histogram_type=@save_histogram_type;
+set use_stat_tables=@save_use_stat_tables;
+
+--echo #
+--echo # Bug mdev-4371: Join with condition supported by index on an empty table  
+--echo #                with optimizer_use_condition_selectivity=3
+--echo #
+
+set use_stat_tables=PREFERABLY; 
+
+CREATE TABLE t1 (a int, b int, INDEX(a));
+
+CREATE TABLE t2 (c int);
+INSERT INTO t2 VALUES (1),(2),(3),(4),(5),(6),(7),(8);
+
+ANALYZE TABLE t1, t2;
+FLUSH TABLES;
+
+set optimizer_use_condition_selectivity=3;
+
+set @save_optimizer_switch=@@optimizer_switch;
+set optimizer_switch='index_condition_pushdown=off';
+
+EXPLAIN EXTENDED
+SELECT * FROM t1, t2 WHERE a > 9;
+SELECT * FROM t1, t2 WHERE a > 9;
+
+set optimizer_switch=@save_optimizer_switch;
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+DROP TABLE t1,t2;
+
+set use_stat_tables=@save_use_stat_tables;
+
+--echo #
+--echo # Bug mdev-4373: condition on a short varchar column   
+--echo #                with optimizer_use_condition_selectivity=3
+--echo #
+
+set use_stat_tables=PREFERABLY;
+
+CREATE TABLE t1 (a varchar(1));
+INSERT INTO t1 VALUES ('x'), ('y');
+
+ANALYZE TABLE t1;
+FLUSH TABLES;
+
+set optimizer_use_condition_selectivity=3;
+
+SELECT * FROM t1 WHERE a <= 'w';
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+DROP TABLE t1;
+
+set use_stat_tables=@save_use_stat_tables;
+
+--echo #
+--echo # Bug mdev-4372: exists subquery in WHERE   
+--echo #                with optimizer_use_condition_selectivity=3
+--echo #
+
+set use_stat_tables = PREFERABLY;
+
+CREATE TABLE t1 (a int);
+INSERT INTO t1 VALUES 
+  (1),(7),(4),(7),(0),(2),(9),(4),(0),(9),(1),(3),(8),(8);
+CREATE TABLE t2 (b int);
+INSERT INTO t2 VALUES (4),(5),(2),(5),(1),(1),(2);
+
+ANALYZE TABLE t1, t2;
+FLUSH TABLES;
+
+set optimizer_use_condition_selectivity=3;
+
+EXPLAIN EXTENDED
+SELECT * FROM t1, t2 WHERE EXISTS ( SELECT 1 FROM t1, t2 ) AND a != b OR a <= 4;
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+DROP TABLE t1,t2;
+
+set use_stat_tables=@save_use_stat_tables;
+
+
+--echo #
+--echo # Bug mdev-4363: selectivity of the condition a IS NULL OR IS NOT NULL   
+--echo #                with optimizer_use_condition_selectivity=3
+--echo #
+
+set use_stat_tables = PREFERABLY;
+
+CREATE TABLE t1 (a int);
+INSERT INTO t1 VALUES 
+  (1),(7),(4),(7),(NULL),(2),(NULL),(4),(NULL),(NULL),(1),(3),(8),(8);
+
+ANALYZE TABLE t1;
+FLUSH TABLE t1;
+
+set optimizer_use_condition_selectivity=3;
+
+EXPLAIN EXTENDED
+SELECT * FROM t1 WHERE a IS NULL;
+EXPLAIN EXTENDED
+SELECT * FROM t1 WHERE a IS NOT NULL;
+EXPLAIN EXTENDED
+SELECT * FROM t1 WHERE a IS NULL OR a IS NOT NULL;
+EXPLAIN EXTENDED
+SELECT * FROM t1 WHERE a IS NULL OR a < 5;
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+DROP TABLE t1;
+
+set use_stat_tables=@save_use_stat_tables;
+
+--echo #
+--echo # Bug mdev-4378: 2-way join with a materialized IN subquery in WHERE   
+--echo #                when optimizer_use_condition_selectivity=4
+--echo #
+
+set use_stat_tables=PREFERABLY;
+set histogram_size=50;
+set histogram_type=SINGLE_PREC_HB;
+
+CREATE TABLE t1 (a INT) ENGINE=MyISAM;
+INSERT INTO t1 VALUES (8),(9),(6);
+
+CREATE TABLE t2 (b INT) ENGINE=MyISAM;
+INSERT INTO t2 VALUES (8),(1),(8),(9),(24),(6),(1),(6),(2),(4);
+
+CREATE TABLE t3 (ln VARCHAR(16)) ENGINE=MyISAM;
+INSERT INTO t3 VALUES 
+('smith'),('black'),('white'),('jones'),
+('brown'),('taylor'),('anderson'),('taylor');
+
+ANALYZE TABLE t1, t2, t3;
+FLUSH TABLES;
+
+set  optimizer_use_condition_selectivity=4;
+
+SELECT * FROM t1, t2 WHERE 'garcia' IN ( SELECT MIN( ln ) FROM t3 WHERE ln = 'sun' );
+
+set histogram_size=@save_histogram_size;
+set histogram_type=@save_histogram_type;
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+DROP TABLE t1,t2,t3;
+
+set use_stat_tables=@save_use_stat_tables;
+
+--echo #
+--echo # Bug mdev-4380: 2-way join with a materialized IN subquery in WHERE   
+--echo #                when optimizer_use_condition_selectivity=3
+--echo #
+
+set  use_stat_tables=PREFERABLY;
+CREATE TABLE t1 (a INT);
+INSERT INTO t1 VALUES (5),(9);
+
+CREATE TABLE t2 (b VARCHAR(8));
+INSERT INTO t2 VALUES ('red'),('blue');
+
+CREATE TABLE t3 (c VARCHAR(8), d VARCHAR(8));
+INSERT INTO t3 VALUES ('white','black'),('cyan','yellow');
+
+ANALYZE TABLE t1, t2, t3;
+FLUSH TABLES;
+
+set optimizer_use_condition_selectivity=3;
+
+SELECT * FROM t1, t2 WHERE ( 'orange', 'green' ) IN ( 
+  SELECT MAX(c), MAX(d) FROM t3, t2 WHERE c >= d AND b = c 
+);
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+DROP TABLE t1,t2,t3;
+
+set use_stat_tables=@save_use_stat_tables;
+
+--echo #
+--echo # Bug mdev-4389: join with degenerated range condition in WHERE   
+--echo #                when optimizer_use_condition_selectivity=3
+--echo #
+
+set use_stat_tables=PREFERABLY;
+
+CREATE TABLE t1 (f1 VARCHAR(1));
+INSERT t1 VALUES ('p'),('q');
+
+CREATE TABLE t2 (f2 VARCHAR(1));
+INSERT INTO t2 VALUES 
+  ('o'),('v'),('f'),('f'),('e'),('l'),('j'),('p'),('r'),('j'),
+  ('j'),('u'),('i'),('r'),('x'),('a'),('x'),('s');
+
+ANALYZE TABLE t1, t2;
+FLUSH TABLES;
+
+SET optimizer_use_condition_selectivity=3;
+
+SELECT * FROM t1, t2 AS t2a, t2 AS t2b WHERE f1 <= 'a' AND t2a.f2 = f1;
+
+set optimizer_use_condition_selectivity=@save_optimizer_use_condition_selectivity;
+
+DROP TABLE t1,t2;
+
+set use_stat_tables=@save_use_stat_tables;
+
+
+set use_stat_tables=@save_use_stat_tables;

mysql-test/t/selectivity_innodb.test

@@ -0,0 +1,12 @@
+--source include/have_innodb.inc
+
+SET SESSION STORAGE_ENGINE='InnoDB';
+
+set @save_optimizer_switch_for_selectivity_test=@@optimizer_switch;
+set optimizer_switch='extended_keys=on';
+
+--source selectivity.test
+
+set optimizer_switch=@save_optimizer_switch_for_selectivity_test;
+
+SET SESSION STORAGE_ENGINE=DEFAULT;

mysql-test/t/statistics.test

@@ -165,6 +165,37 @@ SELECT
      WHERE t1.e IS NOT NULL AND t1.b IS NOT NULL AND t1.d IS NOT NULL)
   AS 'ARITY 3';
 
+
+DELETE FROM mysql.column_stats;
+
+set histogram_size=4;
+ANALYZE TABLE t1;
+
+SELECT db_name, table_name, column_name,
+       min_value, max_value,
+       nulls_ratio, avg_frequency,
+       hist_size, hist_type, HEX(histogram)
+  FROM mysql.column_stats;
+
+DELETE FROM mysql.column_stats;
+
+set histogram_size=8;
+set histogram_type='DOUBLE_PREC_HB';
+ANALYZE TABLE t1;
+
+SELECT db_name, table_name, column_name,
+       min_value, max_value,
+       nulls_ratio, avg_frequency,
+       hist_size, hist_type, HEX(histogram)
+  FROM mysql.column_stats;
+
+DELETE FROM mysql.column_stats;
+
+set histogram_size= 0;
+set histogram_type=default;
+ANALYZE TABLE t1;
+
+
 CREATE TABLE t3 (
   a int NOT NULL PRIMARY KEY,
   b varchar(32),
@@ -559,6 +590,96 @@ DELETE FROM mysql.table_stats;
 DELETE FROM mysql.column_stats;
 DELETE FROM mysql.index_stats;
   
+--echo #
+--echo # Bug mdev-4357: empty string as a value of the HIST_SIZE column 
+--echo #                from mysql.column_stats
+--echo #
+
+create table t1 (a int);
+insert into t1 values (1),(2),(3);
+
+set histogram_size=10;
+
+analyze table t1  persistent for all;
+
+select db_name, table_name, column_name,
+       min_value, max_value,
+       nulls_ratio, avg_frequency,
+       hist_size, hist_type, HEX(histogram)
+  FROM mysql.column_stats;
+
+set histogram_size=default;
+
+drop table t1;
+
+--echo #
+--echo # Bug mdev-4359: wrong setting of the HIST_SIZE column 
+--echo # (see also mdev-4357)         from mysql.column_stats
+--echo #
+
+create table t1 ( a int);
+insert into t1 values (1),(2),(3),(4),(5);
+
+set histogram_size=10;
+set histogram_type='double_prec_hb';
+
+show variables like 'histogram%';
+
+analyze table t1 persistent for all;
+
+select db_name, table_name, column_name,
+       min_value, max_value,
+       nulls_ratio, avg_frequency,
+       hist_size, hist_type, HEX(histogram)
+  FROM mysql.column_stats;
+
+set histogram_size=default;
+set histogram_type=default;
+
+drop table t1;
+
+--echo #
+--echo # Bug mdev-4369: histogram for a column with many distinct values 
+--echo #
+
+
+CREATE TABLE t1 (id int);
+CREATE TABLE t2 (id int);
+
+INSERT INTO t1 (id) VALUES (1), (1), (1),(1);
+INSERT INTO t1 (id) SELECT id FROM t1;
+INSERT INTO t1 SELECT id+1 FROM t1;
+INSERT INTO t1 SELECT id+2 FROM t1;
+INSERT INTO t1 SELECT id+4 FROM t1;
+INSERT INTO t1 SELECT id+8 FROM t1;
+INSERT INTO t1 SELECT id+16 FROM t1;
+INSERT INTO t1 SELECT id+32 FROM t1;
+INSERT INTO t1 SELECT id+64 FROM t1;
+INSERT INTO t1 SELECT id+128 FROM t1;
+INSERT INTO t1 SELECT id+256 FROM t1;
+INSERT INTO t1 SELECT id+512 FROM t1;
+
+INSERT INTO t2 SELECT id FROM t1 ORDER BY id*rand();
+
+SELECT COUNT(*) FROM t2;
+SELECT COUNT(DISTINCT id) FROM t2;
+
+set @@tmp_table_size=1024*16;
+set @@max_heap_table_size=1024*16;
+
+set histogram_size=63;
+
+analyze table t2 persistent for all;
+
+select db_name, table_name, column_name,
+       min_value, max_value,
+       nulls_ratio, avg_frequency,
+       hist_size, hist_type, HEX(histogram)
+  FROM mysql.column_stats;
+
+set histogram_size=default;
+
+drop table t1, t2;
+
 set use_stat_tables=@save_use_stat_tables;
 
-    

scripts/mysql_system_tables.sql

@@ -114,6 +114,6 @@ set @had_proxies_priv_table= @@warning_count != 0;
 
 CREATE TABLE IF NOT EXISTS table_stats (db_name varchar(64) NOT NULL, table_name varchar(64) NOT NULL, cardinality bigint(21) unsigned DEFAULT NULL, PRIMARY KEY (db_name,table_name) ) ENGINE=MyISAM CHARACTER SET utf8 COLLATE utf8_bin comment='Statistics on Tables';
 
-CREATE TABLE IF NOT EXISTS column_stats (db_name varchar(64) NOT NULL, table_name varchar(64) NOT NULL, column_name varchar(64) NOT NULL, min_value varchar(255) DEFAULT NULL, max_value varchar(255) DEFAULT NULL, nulls_ratio decimal(12,4) DEFAULT NULL, avg_length decimal(12,4) DEFAULT NULL, avg_frequency decimal(12,4) DEFAULT NULL, PRIMARY KEY (db_name,table_name,column_name) ) ENGINE=MyISAM CHARACTER SET utf8 COLLATE utf8_bin comment='Statistics on Columns';
+CREATE TABLE IF NOT EXISTS column_stats (db_name varchar(64) NOT NULL, table_name varchar(64) NOT NULL, column_name varchar(64) NOT NULL, min_value varchar(255) DEFAULT NULL, max_value varchar(255) DEFAULT NULL, nulls_ratio decimal(12,4) DEFAULT NULL, avg_length decimal(12,4) DEFAULT NULL, avg_frequency decimal(12,4) DEFAULT NULL, hist_size tinyint unsigned, hist_type enum('SINGLE_PREC_HB','DOUBLE_PREC_HB'), histogram varbinary(255), PRIMARY KEY (db_name,table_name,column_name) ) ENGINE=MyISAM CHARACTER SET utf8 COLLATE utf8_bin comment='Statistics on Columns';
 
 CREATE TABLE IF NOT EXISTS index_stats (db_name varchar(64) NOT NULL, table_name varchar(64) NOT NULL, index_name varchar(64) NOT NULL, prefix_arity int(11) unsigned NOT NULL, avg_frequency decimal(12,4) DEFAULT NULL, PRIMARY KEY (db_name,table_name,index_name,prefix_arity) ) ENGINE=MyISAM CHARACTER SET utf8 COLLATE utf8_bin comment='Statistics on Indexes';

sql/field.cc

@@ -1273,6 +1273,37 @@ out_of_range:
   return 1;
 }
 
+
+/**
+  @brief
+  Determine the relative position of the field value in a numeric interval
+
+  @details
+  The function returns a double number between 0.0 and 1.0 as the relative
+  position of the value of the this field in the numeric interval of [min,max].
+  If the value is not in the interval the the function returns 0.0 when
+  the value is less than min, and, 1.0 when the value is greater than max.
+
+  @param  min  value of the left end of the interval
+  @param  max  value of the right end of the interval
+
+  @return
+  relative position of the field value in the numeric interval [min,max] 
+*/
+
+double Field_num::pos_in_interval(Field *min, Field *max)
+{
+  double n, d;
+  n= val_real() - min->val_real();
+  if (n < 0)
+    return 0.0;
+  d= max->val_real() - min->val_real();
+  if (d <= 0)
+    return 1.0;
+  return min(n/d, 1.0);
+}
+
+
 /**
   Process decimal library return codes and issue warnings for overflow and
   truncation.
@@ -1344,6 +1375,8 @@ Field::Field(uchar *ptr_arg,uint32 length_arg,uchar *null_ptr_arg,
   comment.length=0;
   field_index= 0;   
   is_stat_field= FALSE;
+  cond_selectivity= 1.0;
+  next_equal_field= NULL;
 }
 
 
@@ -6167,6 +6200,79 @@ int Field_str::store(double nr)
   return store(buff, length, &my_charset_numeric);
 }
 
+static
+inline ulonglong char_prefix_to_ulonglong(uchar *src)
+{
+  uint sz= sizeof(ulonglong);
+  for (uint i= 0; i < sz/2; i++)
+  {
+    uchar tmp= src[i];
+    src[i]= src[sz-1-i];
+    src[sz-1-i]= tmp;
+  }
+  return uint8korr(src); 
+}
+
+/**
+  @brief
+  Determine the relative position of the field value in a string interval
+
+  @details
+  The function returns a double number between 0.0 and 1.0 as the relative
+  position of the value of the this field in the string interval of [min,max].
+  If the value is not in the interval the the function returns 0.0 when
+  the value is less than min, and, 1.0 when the value is greater than max.
+
+  @note
+  To calculate the relative position of the string value v in the interval
+  [min, max] the function first converts the beginning of these three
+  strings v, min, max into the strings that are used for byte comparison.
+  For each string not more sizeof(ulonglong) first bytes are taken
+  from the result of conversion. Then these bytes are interpreted as the
+  big-endian representation of an ulonglong integer. The values of these
+  integer numbers obtained for the strings v, min, max are used to calculate
+  the position of v in [min,max] in the same way is it's done for numeric
+  fields (see Field_num::pos_in_interval).
+
+  @todo
+  Improve the procedure for the case when min and max have the same
+  beginning
+     
+  @param  min  value of the left end of the interval
+  @param  max  value of the right end of the interval
+
+  @return
+  relative position of the field value in the string interval [min,max] 
+*/
+
+double Field_str::pos_in_interval(Field *min, Field *max)
+{
+  uchar mp_prefix[sizeof(ulonglong)];
+  uchar minp_prefix[sizeof(ulonglong)];
+  uchar maxp_prefix[sizeof(ulonglong)];
+  ulonglong mp, minp, maxp;
+  my_strnxfrm(charset(), mp_prefix, sizeof(mp),
+              ptr + length_size(),
+              data_length());
+  my_strnxfrm(charset(), minp_prefix, sizeof(minp),
+              min->ptr + length_size(),
+              min->data_length());
+  my_strnxfrm(charset(), maxp_prefix, sizeof(maxp),
+              max->ptr + length_size(),
+              max->data_length());
+  mp= char_prefix_to_ulonglong(mp_prefix);
+  minp= char_prefix_to_ulonglong(minp_prefix);
+  maxp= char_prefix_to_ulonglong(maxp_prefix);
+  double n, d;
+  n= mp - minp;
+  if (n < 0)
+    return 0.0;
+  d= maxp - minp;
+  if (d <= 0)
+    return 1.0;
+  return min(n/d, 1.0);
+}
+
 
 uint Field::is_equal(Create_field *new_field)
 {
@@ -8378,6 +8484,36 @@ my_decimal *Field_bit::val_decimal(my_decimal *deciaml_value)
 }
 
 
+/**
+  @brief
+  Determine the relative position of the field value in a bit interval
+
+  @details
+  The function returns a double number between 0.0 and 1.0 as the relative
+  position of the value of the this field in the bit interval of [min,max].
+  If the value is not in the interval the the function returns 0.0 when
+  the value is less than min, and, 1.0 when the value is greater than max.
+
+  @param  min  value of the left end of the interval
+  @param  max  value of the right end of the interval
+
+  @return
+  relative position of the field value in the bit interval [min,max] 
+*/
+
+double Field_bit::pos_in_interval(Field *min, Field *max)
+{
+  double n, d;
+  n= val_real() - min->val_real();
+  if (n < 0)
+    return 0.0;
+  d= max->val_real() - min->val_real();
+  if (d <= 0)
+    return 1.0;
+  return min(n/d, 1.0);
+}
+
+
 /*
   Compare two bit fields using pointers within the record.
   SYNOPSIS

sql/field.h

@@ -220,7 +220,23 @@ public:
    */
   bool is_created_from_null_item;
 
-  bool is_stat_field; /* TRUE in Field objects created for column min/max values */
+  /* TRUE in Field objects created for column min/max values */
+  bool is_stat_field; 
+
+  /* 
+    Selectivity of the range condition over this field.
+    When calculating this selectivity a range predicate
+    is taken into account only if:
+    - it is extracted from the WHERE clause
+    - it depends only on the table the field belongs to 
+  */
+  double cond_selectivity;
+
+  /* 
+    The next field in the class of equal fields at the top AND level
+    of the WHERE clause
+  */ 
+  Field *next_equal_field;
 
   /*
     This structure is used for statistical data on the column
@@ -456,6 +472,10 @@ public:
     }
     return update_fl;
   }
+  virtual void store_field_value(uchar *val, uint len)
+  {
+     memcpy(ptr, val, len);
+  }
   virtual uint decimals() const { return 0; }
   /*
     Caller beware: sql_type can change str.Ptr, so check
@@ -703,6 +723,12 @@ public:
   virtual bool hash_join_is_possible() { return TRUE; }
   virtual bool eq_cmp_as_binary() { return TRUE; }
 
+  /* Position of the field value within the interval of [min, max] */
+  virtual double pos_in_interval(Field *min, Field *max)
+  {
+    return (double) 0.5; 
+  }
+
   friend int cre_myisam(char * name, register TABLE *form, uint options,
 			ulonglong auto_increment_value);
   friend class Copy_field;
@@ -821,6 +847,7 @@ public:
   bool get_int(CHARSET_INFO *cs, const char *from, uint len, 
                longlong *rnd, ulonglong unsigned_max, 
                longlong signed_min, longlong signed_max);
+  double pos_in_interval(Field *min, Field *max);
 };
 
 
@@ -866,6 +893,8 @@ public:
   virtual bool str_needs_quotes() { return TRUE; }
   uint is_equal(Create_field *new_field);
   bool eq_cmp_as_binary() { return test(flags & BINARY_FLAG); }
+  virtual uint length_size() { return 0; }
+  double pos_in_interval(Field *min, Field *max);
 };
 
 /* base class for Field_string, Field_varstring and Field_blob */
@@ -1894,6 +1923,7 @@ public:
                        uint new_null_bit);
   uint is_equal(Create_field *new_field);
   void hash(ulong *nr, ulong *nr2);
+  uint length_size() { return length_bytes; }
 private:
   int do_save_field_metadata(uchar *first_byte);
 };
@@ -2275,6 +2305,11 @@ public:
     }
     return update_fl;
   }
+  void store_field_value(uchar *val, uint len)
+  {
+    store(*((longlong *)val), TRUE);
+  }
+  double pos_in_interval(Field *min, Field *max);
   void get_image(uchar *buff, uint length, CHARSET_INFO *cs)
   { get_key_image(buff, length, itRAW); }   
   void set_image(const uchar *buff,uint length, CHARSET_INFO *cs)

sql/item_cmpfunc.cc

@@ -5568,7 +5568,8 @@ Item *Item_bool_rowready_func2::negated_item()
 */
 
 Item_equal::Item_equal(Item *f1, Item *f2, bool with_const_item)
-  : Item_bool_func(), eval_item(0), cond_false(0), context_field(NULL)
+  : Item_bool_func(), eval_item(0), cond_false(0), context_field(NULL),
+    link_equal_fields(FALSE)
 {
   const_item_cache= 0;
   with_const= with_const_item;
@@ -5592,7 +5593,8 @@ Item_equal::Item_equal(Item *f1, Item *f2, bool with_const_item)
 */
 
 Item_equal::Item_equal(Item_equal *item_equal)
-  : Item_bool_func(), eval_item(0), cond_false(0), context_field(NULL)
+  : Item_bool_func(), eval_item(0), cond_false(0), context_field(NULL),
+    link_equal_fields(FALSE)
 {
   const_item_cache= 0;
   List_iterator_fast<Item> li(item_equal->equal_items);
@@ -5918,6 +5920,9 @@ bool Item_equal::fix_fields(THD *thd, Item **ref)
   DBUG_ASSERT(fixed == 0);
   Item_equal_fields_iterator it(*this);
   Item *item;
+  Field *first_equal_field;
+  Field *last_equal_field;
+  Field *prev_equal_field= NULL;
   not_null_tables_cache= used_tables_cache= 0;
   const_item_cache= 0;
   while ((item= it++))
@@ -5931,7 +5936,18 @@ bool Item_equal::fix_fields(THD *thd, Item **ref)
       maybe_null= 1;
     if (!item->get_item_equal())
       item->set_item_equal(this);
+    if (link_equal_fields && item->real_item()->type() == FIELD_ITEM)
+    {
+      last_equal_field= ((Item_field *) (item->real_item()))->field;
+      if (!prev_equal_field)
+        first_equal_field= last_equal_field;
+      else
+        prev_equal_field->next_equal_field= last_equal_field;
+      prev_equal_field= last_equal_field;         
+    }
   }
+  if (prev_equal_field && last_equal_field != first_equal_field)
+    last_equal_field->next_equal_field= first_equal_field;
   fix_length_and_dec();
   fixed= 1;
   return FALSE;

sql/item_cmpfunc.h

@@ -1737,6 +1737,8 @@ class Item_equal: public Item_bool_func
   */
   Item_field *context_field;
 
+  bool link_equal_fields;
+
 public:
 
   COND_EQUAL *upper_levels;       /* multiple equalities of upper and levels */
@@ -1774,6 +1776,8 @@ public:
   CHARSET_INFO *compare_collation();
 
   void set_context_field(Item_field *ctx_field) { context_field= ctx_field; }
+  void set_link_equal_fields(bool flag) { link_equal_fields= flag; }
+  friend class Item_equal_fields_iterator;
   friend class Item_equal_iterator<List_iterator_fast,Item>;
   friend class Item_equal_iterator<List_iterator,Item>;
   friend Item *eliminate_item_equal(COND *cond, COND_EQUAL *upper_levels,

sql/opt_range.cc

@@ -117,6 +117,7 @@
 #include "records.h"          // init_read_record, end_read_record
 #include <m_ctype.h>
 #include "sql_select.h"
+#include "sql_statistics.h"
 #include "filesort.h"         // filesort_free_buffers
 
 #ifndef EXTRA_DEBUG
@@ -3217,6 +3218,327 @@ int SQL_SELECT::test_quick_select(THD *thd, key_map keys_to_use,
   DBUG_RETURN(records ? test(quick) : -1);
 }
 
+/****************************************************************************
+ * Condition selectivity module
+ ****************************************************************************/
+
+
+/*
+  Build descriptors of pseudo-indexes over columns to perform range analysis
+
+  SYNOPSIS
+    create_key_parts_for_pseudo_indexes()
+      param       IN/OUT data structure for the descriptors to be built 
+      used_fields bitmap of columns for which the descriptors are to be built          
+
+  DESCRIPTION
+    For each column marked in the bitmap used_fields the function builds
+    a descriptor of a single-component pseudo-index over this column that
+    can be used for the range analysis of the predicates over this columns. 
+    The descriptors are created in the memory of param->mem_root. 
+   
+  RETURN
+    FALSE  in the case of success
+    TRUE   otherwise
+*/
+
+static
+bool create_key_parts_for_pseudo_indexes(RANGE_OPT_PARAM *param,
+                                         MY_BITMAP *used_fields)
+{
+  Field **field_ptr;
+  TABLE *table= param->table;
+  uint parts= 0;
+
+  for (field_ptr= table->field; *field_ptr; field_ptr++)
+  {
+    if (bitmap_is_set(used_fields, (*field_ptr)->field_index))
+      parts++;
+  }
+
+  KEY_PART *key_part;
+  uint keys= 0;
+
+  if (!(key_part= (KEY_PART *)  alloc_root(param->mem_root,
+                                           sizeof(KEY_PART) * parts)))
+    return TRUE;
+
+  param->key_parts= key_part;
+
+  for (field_ptr= table->field; *field_ptr; field_ptr++)
+  {
+    if (bitmap_is_set(used_fields, (*field_ptr)->field_index))
+    {
+      Field *field= *field_ptr;
+      uint16 store_length;
+      key_part->key= keys;
+      key_part->part= 0;
+      key_part->length= (uint16) field->key_length();
+      store_length= key_part->length;
+      if (field->real_maybe_null())
+        store_length+= HA_KEY_NULL_LENGTH;
+      if (field->real_type() == MYSQL_TYPE_VARCHAR)
+        store_length+= HA_KEY_BLOB_LENGTH;
+      key_part->store_length= store_length; 
+      key_part->field= field; 
+      key_part->image_type= Field::itRAW;
+      key_part->flag= 0;
+      param->key[keys]= key_part;
+      keys++;
+      key_part++;
+    }
+  }
+  param->keys= keys;
+  param->key_parts_end= key_part;
+
+  return FALSE;
+}
+
+
+/*
+  Estimate the number of rows in all ranges built for a column
+  by the range optimizer  
+
+  SYNOPSIS
+    records_in_column_ranges()
+      param      the data structure to access descriptors of pseudo indexes
+                 built over columns used in the condition of the processed query
+      idx        the index of the descriptor of interest in param
+      tree       the tree representing ranges built for the interesting column         
+
+  DESCRIPTION
+    This function retrieves the ranges represented by the SEL_ARG 'tree' and
+    for each of them r it calls the function get_column_range_cardinality()
+    that estimates the number of expected rows in r. It is assumed that param
+    is the data structure containing the descriptors of pseudo-indexes that
+    has been built to perform range analysis of the range conditions imposed
+    on the columns used in the processed query, while idx is the index of the
+    descriptor created in 'param' exactly for the column for which 'tree'
+    has been built by the range optimizer.    
+
+  RETURN
+    the number of rows in the retrieved ranges  
+*/
+
+static
+double records_in_column_ranges(PARAM *param, uint idx, 
+                                SEL_ARG *tree)
+{
+  SEL_ARG_RANGE_SEQ seq;
+  KEY_MULTI_RANGE range;
+  range_seq_t seq_it;
+  double rows;
+  Field *field;
+  uint flags= 0;
+  double total_rows= 0;
+  RANGE_SEQ_IF seq_if = {NULL, sel_arg_range_seq_init, 
+                         sel_arg_range_seq_next, 0, 0};
+  
+  /* Handle cases when we don't have a valid non-empty list of range */
+  if (!tree)
+    return HA_POS_ERROR;
+  if (tree->type == SEL_ARG::IMPOSSIBLE)
+    return (0L);
+
+  field= tree->field;
+
+  seq.keyno= idx;
+  seq.real_keyno= MAX_KEY;
+  seq.param= param;
+  seq.start= tree;
+
+  seq_it= seq_if.init((void *) &seq, 0, flags);
+
+  while (!seq_if.next(seq_it, &range))
+  {
+    key_range *min_endp, *max_endp;
+    min_endp= range.start_key.length? &range.start_key : NULL;
+    max_endp= range.end_key.length? &range.end_key : NULL;
+    rows= get_column_range_cardinality(field, min_endp, max_endp,
+                                       range.range_flag);
+    if (HA_POS_ERROR == rows)
+    {
+      total_rows= HA_POS_ERROR;
+      break;
+    }
+    total_rows += rows;
+  }    
+  return total_rows;
+} 
+
+
+/*
+  Calculate the selectivity of the condition imposed on the rows of a table
+
+  SYNOPSIS
+    calculate_cond_selectivity_for_table()
+      thd        the context handle 
+      table      the table of interest
+      cond       conditions imposed on the rows of the table        
+
+  DESCRIPTION
+    This function calculates the selectivity of range conditions cond imposed
+    on the rows of 'table' in the processed query.
+    The calculated selectivity is assigned to the field table->cond_selectivity.
+    
+  NOTE
+    Currently the selectivities of range conditions over different columns are
+    considered independent. 
+
+  RETURN
+    FALSE  on success
+    TRUE   otherwise 
+*/
+
+bool calculate_cond_selectivity_for_table(THD *thd, TABLE *table, Item *cond)
+{
+  uint keynr;
+  uint max_quick_key_parts= 0;
+  MY_BITMAP *used_fields= &table->cond_set;
+  double table_records= table->stat_records(); 
+  DBUG_ENTER("calculate_cond_selectivity_for_table");
+
+  table->cond_selectivity= 1.0;
+
+  if (table_records == 0)
+    DBUG_RETURN(FALSE);
+  
+  if (thd->variables.optimizer_use_condition_selectivity > 2 &&
+      !bitmap_is_clear_all(used_fields))
+  {
+    /* 
+      Calculate the selectivity of the range conditions not supported
+      by any index
+    */
+
+    PARAM param;
+    MEM_ROOT alloc;
+    SEL_TREE *tree;
+    SEL_ARG **key, **end;
+    double rows;
+    uint idx= 0;
+  
+    init_sql_alloc(&alloc, thd->variables.range_alloc_block_size, 0,
+                   MYF(MY_THREAD_SPECIFIC));
+    param.thd= thd;
+    param.mem_root= &alloc;
+    param.old_root= thd->mem_root;
+    param.table= table;
+    param.is_ror_scan= FALSE;
+
+    if (create_key_parts_for_pseudo_indexes(&param, used_fields))
+      goto free_alloc;
+
+    param.prev_tables= param.read_tables= 0;
+    param.current_table= table->map;
+    param.using_real_indexes= FALSE;
+    param.real_keynr[0]= 0;
+    param.alloced_sel_args= 0;
+
+    thd->no_errors=1;		    
+
+    tree= get_mm_tree(&param, cond);
+
+    if (!tree)
+      goto free_alloc;
+    
+    table->reginfo.impossible_range= 0;
+    if (tree->type == SEL_TREE::IMPOSSIBLE)
+    {
+      rows= 0;
+      table->reginfo.impossible_range= 1;
+      goto free_alloc;
+    }  
+    else if (tree->type == SEL_TREE::MAYBE)
+    {
+      rows= table_records;
+      goto free_alloc;
+    }        
+
+    for (key= tree->keys, end= key + param.keys; key != end; key++, idx++)
+    {
+      if (*key)
+      {
+        if ((*key)->type == SEL_ARG::IMPOSSIBLE)
+	{
+          rows= 0;
+          table->reginfo.impossible_range= 1;
+          goto free_alloc;
+        }          
+        else
+        {
+          rows= records_in_column_ranges(&param, idx, *key);
+          if (rows != HA_POS_ERROR)
+            (*key)->field->cond_selectivity= rows/table_records;
+        } 
+      }
+    }
+
+    for (Field **field_ptr= table->field; *field_ptr; field_ptr++)
+    {
+      Field *table_field= *field_ptr;   
+      if (bitmap_is_set(table->read_set, table_field->field_index) &&
+          table_field->cond_selectivity < 1.0)
+        table->cond_selectivity*= table_field->cond_selectivity;
+    }
+
+  free_alloc:
+    thd->mem_root= param.old_root;
+    free_root(&alloc, MYF(0));
+
+  }
+
+  /* Calculate the selectivity of the range conditions supported by indexes */
+
+  bitmap_clear_all(used_fields);
+
+  for (keynr= 0;  keynr < table->s->keys; keynr++)
+  {
+    if (table->quick_keys.is_set(keynr))
+      set_if_bigger(max_quick_key_parts, table->quick_key_parts[keynr]);
+  }
+
+  for (uint quick_key_parts= max_quick_key_parts;
+       quick_key_parts; quick_key_parts--)
+  {
+    for (keynr= 0;  keynr < table->s->keys; keynr++)
+    {
+      if (table->quick_keys.is_set(keynr) &&
+          table->quick_key_parts[keynr] == quick_key_parts)
+      {
+        uint i;
+        uint used_key_parts= table->quick_key_parts[keynr];
+        double quick_cond_selectivity= table->quick_rows[keynr] / 
+	                               table_records;
+        KEY *key_info= table->key_info + keynr;
+        KEY_PART_INFO* key_part= key_info->key_part;
+        for (i= 0; i < used_key_parts; i++, key_part++)
+        {
+          if (bitmap_is_set(used_fields, key_part->fieldnr-1))
+	    break; 
+          bitmap_set_bit(used_fields, key_part->fieldnr-1);
+        }
+        if (i)
+        {
+          table->cond_selectivity*= quick_cond_selectivity;
+          if (i != used_key_parts)
+	  {
+            double f1= key_info->actual_rec_per_key(i-1);
+            double f2= key_info->actual_rec_per_key(i);
+            table->cond_selectivity*= f1 / f2;
+          }
+        }
+      } 
+    }
+  } 
+
+  DBUG_RETURN(FALSE);
+}
+
+/****************************************************************************
+ * Condition selectivity code ends
+ ****************************************************************************/
+
 /****************************************************************************
  * Partition pruning module
  ****************************************************************************/

sql/opt_range.h

@@ -1042,6 +1042,8 @@ SQL_SELECT *make_select(TABLE *head, table_map const_tables,
 			table_map read_tables, COND *conds,
                         bool allow_null_cond,  int *error);
 
+bool calculate_cond_selectivity_for_table(THD *thd, TABLE *table, Item *cond);
+
 #ifdef WITH_PARTITION_STORAGE_ENGINE
 bool prune_partitions(THD *thd, TABLE *table, Item *pprune_cond);
 void store_key_image_to_rec(Field *field, uchar *ptr, uint len);

sql/opt_range_mrr.cc

@@ -268,8 +268,10 @@ walk_up_n_right:
     range->end_key.keypart_map= make_prev_keypart_map(cur->max_key_parts);
 
     if (!(cur->min_key_flag & ~NULL_RANGE) && !cur->max_key_flag &&
-        (uint)key_tree->part+1 == seq->param->table->key_info[seq->real_keyno].key_parts &&
-        (seq->param->table->key_info[seq->real_keyno].flags & HA_NOSAME) &&
+        (seq->real_keyno == MAX_KEY ||
+         ((uint)key_tree->part+1 ==
+          seq->param->table->key_info[seq->real_keyno].key_parts &&
+	  (seq->param->table->key_info[seq->real_keyno].flags & HA_NOSAME))) &&
         range->start_key.length == range->end_key.length &&
         !memcmp(seq->param->min_key,seq->param->max_key,range->start_key.length))
       range->range_flag= UNIQUE_RANGE | (cur->min_key_flag & NULL_RANGE);

sql/opt_subselect.cc

@@ -3908,7 +3908,7 @@ SJ_TMP_TABLE::create_sj_weedout_tmp_table(THD *thd)
                         &tmpname, (uint) strlen(path)+1,
                         &group_buff, (!using_unique_constraint ?
                                       uniq_tuple_length_arg : 0),
-                        &bitmaps, bitmap_buffer_size(1)*3,
+                        &bitmaps, bitmap_buffer_size(1)*5,
                         NullS))
   {
     if (temp_pool_slot != MY_BIT_NONE)

sql/opt_subselect.h

@@ -283,6 +283,7 @@ public:
     {
       pos->records_read=    best_loose_scan_records;
       pos->key=             best_loose_scan_start_key;
+      pos->cond_selectivity= 1.0;
       pos->loosescan_picker.loosescan_key=   best_loose_scan_key;
       pos->loosescan_picker.loosescan_parts= best_max_loose_keypart + 1;
       pos->use_join_buffer= FALSE;

sql/sql_class.h

@@ -504,7 +504,10 @@ typedef struct system_variables
   ulong net_write_timeout;
   ulong optimizer_prune_level;
   ulong optimizer_search_depth;
+  ulong optimizer_use_condition_selectivity;
   ulong use_stat_tables;
+  ulong histogram_size;
+  ulong histogram_type;
   ulong preload_buff_size;
   ulong profiling_history_size;
   ulong read_buff_size;
@@ -4017,6 +4020,7 @@ class Unique :public Sql_alloc
   uint size;
   uint full_size;
   uint min_dupl_count;   /* always 0 for unions, > 0 for intersections */
+  bool with_counters;
 
   bool merge(TABLE *table, uchar *buff, bool without_last_merge);
 

sql/sql_priv.h

@@ -228,6 +228,7 @@ template <class T> bool valid_buffer_range(T jump,
 #define OPTIMIZER_SWITCH_TABLE_ELIMINATION         (1ULL << 26)
 #define OPTIMIZER_SWITCH_EXTENDED_KEYS             (1ULL << 27)
 #define OPTIMIZER_SWITCH_EXISTS_TO_IN              (1ULL << 28)
+#define OPTIMIZER_SWITCH_USE_CONDITION_SELECTIVITY (1ULL << 29)
 
 #define OPTIMIZER_SWITCH_DEFAULT   (OPTIMIZER_SWITCH_INDEX_MERGE | \
                                     OPTIMIZER_SWITCH_INDEX_MERGE_UNION | \

sql/sql_select.cc

@@ -89,12 +89,14 @@ void best_access_path(JOIN *join, JOIN_TAB *s,
                              POSITION *pos, POSITION *loose_scan_pos);
 static void optimize_straight_join(JOIN *join, table_map join_tables);
 static bool greedy_search(JOIN *join, table_map remaining_tables,
-                          uint depth, uint prune_level);
+                          uint depth, uint prune_level,
+                          uint use_cond_selectivity);
 static bool best_extension_by_limited_search(JOIN *join,
                                              table_map remaining_tables,
                                              uint idx, double record_count,
                                              double read_time, uint depth,
-                                             uint prune_level);
+                                             uint prune_level,
+                                             uint use_cond_selectivity);
 static uint determine_search_depth(JOIN* join);
 C_MODE_START
 static int join_tab_cmp(const void *dummy, const void* ptr1, const void* ptr2);
@@ -134,7 +136,8 @@ static COND *build_equal_items(JOIN *join, COND *cond,
                                COND_EQUAL *inherited,
                                List<TABLE_LIST> *join_list,
                                bool ignore_on_conds,
-                               COND_EQUAL **cond_equal_ref);
+                               COND_EQUAL **cond_equal_ref,
+                               bool link_equal_fields= FALSE);
 static COND* substitute_for_best_equal_field(JOIN_TAB *context_tab,
                                              COND *cond,
                                              COND_EQUAL *cond_equal,
@@ -150,8 +153,9 @@ static uint build_bitmap_for_nested_joins(List<TABLE_LIST> *join_list,
 static COND *optimize_cond(JOIN *join, COND *conds,
                            List<TABLE_LIST> *join_list,
                            bool ignore_on_conds,
-			   Item::cond_result *cond_value, 
-                           COND_EQUAL **cond_equal);
+                           Item::cond_result *cond_value, 
+                           COND_EQUAL **cond_equal,
+                           int flags= 0);
 bool const_expression_in_where(COND *conds,Item *item, Item **comp_item);
 static bool create_internal_tmp_table_from_heap2(THD *, TABLE *,
                                      ENGINE_COLUMNDEF *, ENGINE_COLUMNDEF **, 
@@ -279,6 +283,8 @@ enum enum_exec_or_opt {WALK_OPTIMIZATION_TABS , WALK_EXECUTION_TABS};
 JOIN_TAB *first_breadth_first_tab(JOIN *join, enum enum_exec_or_opt tabs_kind);
 JOIN_TAB *next_breadth_first_tab(JOIN *join, enum enum_exec_or_opt tabs_kind,
                                  JOIN_TAB *tab);
+static double table_cond_selectivity(JOIN *join, uint idx, JOIN_TAB *s,
+                                     table_map rem_tables);
 
 #ifndef DBUG_OFF
 
@@ -1158,7 +1164,7 @@ TODO: make view to decide if it is possible to write to WHERE directly or make S
     DBUG_RETURN(1);
 
   conds= optimize_cond(this, conds, join_list, FALSE,
-                       &cond_value, &cond_equal);
+                       &cond_value, &cond_equal, OPT_LINK_EQUAL_FIELDS);
      
   if (thd->is_error())
   {
@@ -3355,6 +3361,7 @@ make_join_statistics(JOIN *join, List<TABLE_LIST> &tables_list,
     table->pos_in_table_list= tables;
     error= tables->fetch_number_of_rows();
     set_statistics_for_table(join->thd, table);
+    bitmap_clear_all(&table->cond_set);
 
 #ifdef WITH_PARTITION_STORAGE_ENGINE
     const bool no_partitions_used= table->no_partitions_used;
@@ -3786,6 +3793,8 @@ make_join_statistics(JOIN *join, List<TABLE_LIST> &tables_list,
       all select distinct fields participate in one index.
     */
     add_group_and_distinct_keys(join, s);
+
+    s->table->cond_selectivity= 1.0;
     
     /*
       Perform range analysis if there are keys it could use (1). 
@@ -3794,7 +3803,8 @@ make_join_statistics(JOIN *join, List<TABLE_LIST> &tables_list,
       Don't do range analysis for materialized subqueries (4).
       Don't do range analysis for materialized derived tables (5)
     */
-    if (!s->const_keys.is_clear_all() &&                            // (1)
+    if ((!s->const_keys.is_clear_all() ||
+	 !bitmap_is_clear_all(&s->table->cond_set)) &&              // (1)
         (!s->table->pos_in_table_list->embedding ||                 // (2)
          (s->table->pos_in_table_list->embedding &&                 // (3)
           s->table->pos_in_table_list->embedding->sj_on_expr)) &&   // (3)
@@ -3802,20 +3812,37 @@ make_join_statistics(JOIN *join, List<TABLE_LIST> &tables_list,
         !(s->table->pos_in_table_list->derived &&                   // (5)
           s->table->pos_in_table_list->is_materialized_derived()))  // (5)
     {
-      ha_rows records;
-      SQL_SELECT *select;
-      select= make_select(s->table, found_const_table_map,
-			  found_const_table_map,
-			  *s->on_expr_ref ? *s->on_expr_ref : conds,
-			  1, &error);
-      if (!select)
-        goto error;
-      records= get_quick_record_count(join->thd, select, s->table,
-				      &s->const_keys, join->row_limit);
-      s->quick=select->quick;
-      s->needed_reg=select->needed_reg;
-      select->quick=0;
-      if (records == 0 && s->table->reginfo.impossible_range)
+      bool impossible_range= FALSE;
+      ha_rows records= HA_POS_ERROR;
+      SQL_SELECT *select= 0;
+      if (!s->const_keys.is_clear_all())
+      {
+        select= make_select(s->table, found_const_table_map,
+			    found_const_table_map,
+			    *s->on_expr_ref ? *s->on_expr_ref : conds,
+			    1, &error);
+        if (!select)
+          goto error;
+        records= get_quick_record_count(join->thd, select, s->table,
+				        &s->const_keys, join->row_limit);
+        s->quick=select->quick;
+        s->needed_reg=select->needed_reg;
+        select->quick=0;
+        impossible_range= records == 0 && s->table->reginfo.impossible_range;
+      }
+      if (!impossible_range)
+      {
+        if (join->thd->variables.optimizer_use_condition_selectivity > 1)
+          calculate_cond_selectivity_for_table(join->thd, s->table, 
+                                               *s->on_expr_ref ?
+                                               *s->on_expr_ref : conds);
+        if (s->table->reginfo.impossible_range)
+	{
+          impossible_range= TRUE;
+          records= 0;
+        }
+      }
+      if (impossible_range)
       {
 	/*
 	  Impossible WHERE or ON expression
@@ -3840,8 +3867,10 @@ make_join_statistics(JOIN *join, List<TABLE_LIST> &tables_list,
 	s->found_records=records;
 	s->read_time= s->quick ? s->quick->read_time : 0.0;
       }
-      delete select;
+      if (select)
+        delete select;
     }
+
   }
 
   if (pull_out_semijoin_tables(join))
@@ -4198,11 +4227,12 @@ add_key_field(JOIN *join,
   else if (!(field->flags & PART_KEY_FLAG))
   {
     // Don't remove column IS NULL on a LEFT JOIN table
-    if (!eq_func || (*value)->type() != Item::NULL_ITEM ||
-        !field->table->maybe_null || field->null_ptr)
-      return;					// Not a key. Skip it
-    optimize= KEY_OPTIMIZE_EXISTS;
-    DBUG_ASSERT(num_values == 1);
+    if (eq_func && (*value)->type() == Item::NULL_ITEM &&
+        field->table->maybe_null && !field->null_ptr)
+    {
+      optimize= KEY_OPTIMIZE_EXISTS;
+      DBUG_ASSERT(num_values == 1);
+    }
   }
   if (optimize != KEY_OPTIMIZE_EXISTS)
   {
@@ -4251,7 +4281,11 @@ add_key_field(JOIN *join,
           break;
       }
       if (is_const)
+      {
         stat[0].const_keys.merge(possible_keys);
+        if (possible_keys.is_clear_all())
+          bitmap_set_bit(&field->table->cond_set, field->field_index);
+      }
       else if (!eq_func)
       {
         /* 
@@ -5322,6 +5356,7 @@ void set_position(JOIN *join,uint idx,JOIN_TAB *table,KEYUSE *key)
   join->positions[idx].table= table;
   join->positions[idx].key=key;
   join->positions[idx].records_read=1.0;	/* This is a const table */
+  join->positions[idx].cond_selectivity= 1.0;
   join->positions[idx].ref_depend_map= 0;
 
 //  join->positions[idx].loosescan_key= MAX_KEY; /* Not a LooseScan */
@@ -6140,6 +6175,8 @@ choose_plan(JOIN *join, table_map join_tables)
 {
   uint search_depth= join->thd->variables.optimizer_search_depth;
   uint prune_level=  join->thd->variables.optimizer_prune_level;
+  uint use_cond_selectivity= 
+         join->thd->variables.optimizer_use_condition_selectivity;
   bool straight_join= test(join->select_options & SELECT_STRAIGHT_JOIN);
   DBUG_ENTER("choose_plan");
 
@@ -6204,7 +6241,8 @@ choose_plan(JOIN *join, table_map join_tables)
       if (search_depth == 0)
         /* Automatically determine a reasonable value for 'search_depth' */
         search_depth= determine_search_depth(join);
-      if (greedy_search(join, join_tables, search_depth, prune_level))
+      if (greedy_search(join, join_tables, search_depth, prune_level,
+                        use_cond_selectivity))
         DBUG_RETURN(TRUE);
     }
   }
@@ -6478,6 +6516,8 @@ optimize_straight_join(JOIN *join, table_map join_tables)
   bool disable_jbuf= join->thd->variables.join_cache_level == 0;
   double    record_count= 1.0;
   double    read_time=    0.0;
+  uint use_cond_selectivity= 
+         join->thd->variables.optimizer_use_condition_selectivity;
   POSITION  loose_scan_pos;
 
   for (JOIN_TAB **pos= join->best_ref + idx ; (s= *pos) ; pos++)
@@ -6494,6 +6534,11 @@ optimize_straight_join(JOIN *join, table_map join_tables)
                      &loose_scan_pos);
 
     join_tables&= ~(s->table->map);
+    double pushdown_cond_selectivity= 1.0;
+    if (use_cond_selectivity > 1)
+      pushdown_cond_selectivity= table_cond_selectivity(join, idx, s,
+                                                        join_tables);
+    join->positions[idx].cond_selectivity= pushdown_cond_selectivity;
     ++idx;
   }
 
@@ -6581,6 +6626,8 @@ optimize_straight_join(JOIN *join, table_map join_tables)
   @param search_depth     controlls the exhaustiveness of the search
   @param prune_level      the pruning heuristics that should be applied during
                           search
+  @param use_cond_selectivity  specifies how the selectivity of the conditions
+                          pushed to a table should be taken into account
 
   @retval
     FALSE       ok
@@ -6592,7 +6639,8 @@ static bool
 greedy_search(JOIN      *join,
               table_map remaining_tables,
               uint      search_depth,
-              uint      prune_level)
+              uint      prune_level,
+              uint      use_cond_selectivity)
 {
   double    record_count= 1.0;
   double    read_time=    0.0;
@@ -6617,7 +6665,8 @@ greedy_search(JOIN      *join,
     /* Find the extension of the current QEP with the lowest cost */
     join->best_read= DBL_MAX;
     if (best_extension_by_limited_search(join, remaining_tables, idx, record_count,
-                                         read_time, search_depth, prune_level))
+                                         read_time, search_depth, prune_level,
+                                         use_cond_selectivity))
       DBUG_RETURN(TRUE);
     /*
       'best_read < DBL_MAX' means that optimizer managed to find
@@ -6856,6 +6905,210 @@ double JOIN::get_examined_rows()
 }
 
 
+static 
+double table_multi_eq_cond_selectivity(JOIN *join, uint idx, JOIN_TAB *s,
+                                       table_map rem_tables, uint keyparts,
+                                       uint16 *ref_keyuse_steps)
+{
+  double sel= 1.0;
+  COND_EQUAL *cond_equal= join->cond_equal;
+
+  if (!cond_equal || !cond_equal->current_level.elements)
+    return sel;
+
+   if (!s->keyuse)
+    return sel;
+
+ Item_equal *item_equal;
+  List_iterator_fast<Item_equal> it(cond_equal->current_level);
+  TABLE *table= s->table;
+  table_map table_bit= table->map;
+  POSITION *pos= &join->positions[idx];
+  
+  while ((item_equal= it++))
+  { 
+    /* 
+      Check whether we need to take into account the selectivity of
+      multiple equality item_equal. If this is the case multiply
+      the current value of sel by this selectivity
+    */
+    table_map used_tables= item_equal->used_tables();
+    if (!(used_tables & table_bit))
+      continue;
+    if (item_equal->get_const())
+      continue;
+
+    Field *fld;
+    bool adjust_sel= FALSE;
+    Item_equal_fields_iterator fi(*item_equal);
+    while((fi++) && !adjust_sel)
+    {
+      Field *fld= fi.get_curr_field();
+      if (fld->table->map != table_bit)
+        continue;
+      if (pos->key == 0)
+        adjust_sel= TRUE;
+      else
+      {
+        uint i;
+        KEYUSE *keyuse= pos->key;
+        uint key= keyuse->key;
+
+        for (i= 0; i < keyparts; i++)
+	{
+          uint fldno;
+          if (is_hash_join_key_no(key))
+	    fldno= keyuse->keypart;
+          else
+            fldno= table->key_info[key].key_part[keyparts-1].fieldnr - 1;        
+          if (fld->field_index == fldno)
+            break;
+        }
+        if (i == keyparts)
+	{
+          /* 
+            Field fld is included in multiple equality item_equal
+            and is not a part of the ref key.
+            The selectivity of the multiple equality must be taken
+            into account unless one of the ref arguments is
+            equal to fld.  
+	  */
+          adjust_sel= TRUE;
+          for (uint j= 0; j < keyparts && adjust_sel; j++)
+	  {
+            if (j > 0)
+              keyuse+= ref_keyuse_steps[j-1];  
+            Item *ref_item= keyuse->val;
+	    if (ref_item->real_item()->type() == Item::FIELD_ITEM)
+	    {
+              Item_field *field_item= (Item_field *) (ref_item->real_item());
+              if (item_equal->contains(field_item->field))
+                adjust_sel= FALSE;              
+	    }
+          }
+        }          
+      }
+    }
+    if (adjust_sel)
+    {
+      /* 
+        If ref == 0 and there are no fields in the multiple equality
+        item_equal that belong to the tables joined prior to s
+        then the selectivity of multiple equality will be set to 1.0.
+      */
+      double eq_fld_sel= 1.0;
+      fi.rewind();
+      while ((fi++))
+      {
+        double curr_eq_fld_sel;
+        fld= fi.get_curr_field();
+        if (!fld->table->map & ~(table_bit | rem_tables))
+          continue;
+        curr_eq_fld_sel= get_column_avg_frequency(fld) /
+                         fld->table->stat_records();
+        if (curr_eq_fld_sel < 1.0)
+          set_if_bigger(eq_fld_sel, curr_eq_fld_sel);
+      }
+      sel*= eq_fld_sel;
+    }
+  } 
+  return sel;
+}
+
+static
+double table_cond_selectivity(JOIN *join, uint idx, JOIN_TAB *s,
+                              table_map rem_tables)
+{
+  uint16 ref_keyuse_steps[MAX_REF_PARTS - 1];
+  Field *field;
+  TABLE *table= s->table;
+  MY_BITMAP *read_set= table->read_set;
+  double sel= s->table->cond_selectivity;
+  double table_records= table->stat_records();
+  POSITION *pos= &join->positions[idx];
+  uint keyparts= 0;
+  uint found_part_ref_or_null= 0;
+
+  /* Discount the selectivity of the access method used to join table s */
+  if (s->quick && s->quick->index != MAX_KEY)
+  {
+    if (pos->key == 0 && table_records > 0)
+    {
+      sel*= table->quick_rows[s->quick->index]/table_records;
+    }
+  }
+  else if (pos->key != 0)
+  {
+    /* A ref/ access or hash join is used to join table */
+    KEYUSE *keyuse= pos->key;
+    KEYUSE *prev_ref_keyuse= keyuse;
+    uint key= keyuse->key;
+    do
+    {
+      if (!(keyuse->used_tables & (rem_tables | table->map)))
+      {
+        if (are_tables_local(s, keyuse->val->used_tables()))
+	{
+          if (is_hash_join_key_no(key))
+	  {
+            if (keyparts == keyuse->keypart)
+              keyparts++;
+          }
+          else
+	  {
+            if (keyparts == keyuse->keypart &&
+                !(~(keyuse->val->used_tables()) & pos->ref_depend_map) &&
+                !(found_part_ref_or_null & keyuse->optimize))
+	    {
+              keyparts++;
+              found_part_ref_or_null|= keyuse->optimize & ~KEY_OPTIMIZE_EQ;
+            }
+          }
+          if (keyparts > keyuse->keypart)
+	  {
+            uint fldno;
+            if (is_hash_join_key_no(key))
+	      fldno= keyuse->keypart;
+            else
+              fldno= table->key_info[key].key_part[keyparts-1].fieldnr - 1;
+            if (keyuse->val->const_item())
+              sel*= table->field[fldno]->cond_selectivity; 
+            if (keyparts > 1)
+	    {
+              ref_keyuse_steps[keyparts-2]= keyuse - prev_ref_keyuse;
+              prev_ref_keyuse= keyuse;
+            }
+          }
+	}
+      }
+      keyuse++;
+    } while (keyuse->table == table && keyuse->key == key);
+  }
+    
+  for (Field **f_ptr=table->field ; (field= *f_ptr) ; f_ptr++)
+  {
+    if (!bitmap_is_set(read_set, field->field_index) ||
+        !field->next_equal_field)
+      continue; 
+    for (Field *next_field= field->next_equal_field; 
+         next_field != field; 
+         next_field= next_field->next_equal_field)
+    {
+      if (!(next_field->table->map & rem_tables) && next_field->table != table)
+      { 
+        sel/= field->cond_selectivity;
+        break;
+      }
+    }
+  }
+
+  sel*= table_multi_eq_cond_selectivity(join, idx, s, rem_tables,
+                                        keyparts, ref_keyuse_steps);
+
+  return sel;
+}
+
+
 /**
   Find a good, possibly optimal, query execution plan (QEP) by a possibly
   exhaustive search.
@@ -6966,6 +7219,8 @@ double JOIN::get_examined_rows()
   @param prune_level      pruning heuristics that should be applied during
                           optimization
                           (values: 0 = EXHAUSTIVE, 1 = PRUNE_BY_TIME_OR_ROWS)
+  @param use_cond_selectivity  specifies how the selectivity of the conditions
+                          pushed to a table should be taken into account
 
   @retval
     FALSE       ok
@@ -6980,7 +7235,8 @@ best_extension_by_limited_search(JOIN      *join,
                                  double    record_count,
                                  double    read_time,
                                  uint      search_depth,
-                                 uint      prune_level)
+                                 uint      prune_level,
+                                 uint      use_cond_selectivity)
 {
   DBUG_ENTER("best_extension_by_limited_search");
 
@@ -7083,16 +7339,25 @@ best_extension_by_limited_search(JOIN      *join,
         }
       }
 
+      double pushdown_cond_selectivity= 1.0;
+      if (use_cond_selectivity > 1)
+        pushdown_cond_selectivity= table_cond_selectivity(join, idx, s,
+				                          remaining_tables &
+                                                          ~real_table_bit);
+      join->positions[idx].cond_selectivity= pushdown_cond_selectivity;
+      double partial_join_cardinality= current_record_count *
+                                        pushdown_cond_selectivity;
       if ( (search_depth > 1) && (remaining_tables & ~real_table_bit) & allowed_tables )
       { /* Recursively expand the current partial plan */
         swap_variables(JOIN_TAB*, join->best_ref[idx], *pos);
         if (best_extension_by_limited_search(join,
                                              remaining_tables & ~real_table_bit,
                                              idx + 1,
-                                             current_record_count,
+                                             partial_join_cardinality,
                                              current_read_time,
                                              search_depth - 1,
-                                             prune_level))
+                                             prune_level,
+                                             use_cond_selectivity))
           DBUG_RETURN(TRUE);
         swap_variables(JOIN_TAB*, join->best_ref[idx], *pos);
       }
@@ -7110,7 +7375,7 @@ best_extension_by_limited_search(JOIN      *join,
         {
           memcpy((uchar*) join->best_positions, (uchar*) join->positions,
                  sizeof(POSITION) * (idx + 1));
-          join->record_count= current_record_count;
+          join->record_count= partial_join_cardinality;
           join->best_read= current_read_time - 0.001;
         }
         DBUG_EXECUTE("opt", print_plan(join, idx+1,
@@ -7755,6 +8020,7 @@ get_best_combination(JOIN *join)
       */
       SJ_MATERIALIZATION_INFO *sjm= cur_pos->table->emb_sj_nest->sj_mat_info;
       j->records= j->records_read= (ha_rows)(sjm->is_sj_scan? sjm->rows : 1);
+      j->cond_selectivity= 1.0;
       JOIN_TAB *jt;
       JOIN_TAB_RANGE *jt_range;
       if (!(jt= (JOIN_TAB*)join->thd->alloc(sizeof(JOIN_TAB)*sjm->tables)) ||
@@ -7818,6 +8084,7 @@ get_best_combination(JOIN *join)
       to access join->best_positions[]. 
     */
     j->records_read= (ha_rows)join->best_positions[tablenr].records_read;
+    j->cond_selectivity= join->best_positions[tablenr].cond_selectivity;
     join->map2table[j->table->tablenr]= j;
 
     /* If we've reached the end of sjm nest, switch back to main sequence */
@@ -11796,7 +12063,8 @@ static bool check_equality(THD *thd, Item *item, COND_EQUAL *cond_equal,
 */
 
 static COND *build_equal_items_for_cond(THD *thd, COND *cond,
-                                        COND_EQUAL *inherited)
+                                        COND_EQUAL *inherited,
+                                        bool link_item_fields)
 {
   Item_equal *item_equal;
   COND_EQUAL cond_equal;
@@ -11843,6 +12111,7 @@ static COND *build_equal_items_for_cond(THD *thd, COND *cond,
       List_iterator_fast<Item_equal> it(cond_equal.current_level);
       while ((item_equal= it++))
       {
+        item_equal->set_link_equal_fields(link_item_fields);
         item_equal->fix_fields(thd, NULL);
         item_equal->update_used_tables();
         set_if_bigger(thd->lex->current_select->max_equal_elems,
@@ -11862,7 +12131,8 @@ static COND *build_equal_items_for_cond(THD *thd, COND *cond,
     while ((item= li++))
     { 
       Item *new_item;
-      if ((new_item= build_equal_items_for_cond(thd, item, inherited)) != item)
+      if ((new_item= build_equal_items_for_cond(thd, item, inherited, FALSE))
+          != item)
       {
         /* This replacement happens only for standalone equalities */
         /*
@@ -12025,14 +12295,15 @@ static COND *build_equal_items(JOIN *join, COND *cond,
                                COND_EQUAL *inherited,
                                List<TABLE_LIST> *join_list,
                                bool ignore_on_conds,
-                               COND_EQUAL **cond_equal_ref)
+                               COND_EQUAL **cond_equal_ref,
+                               bool link_equal_fields)
 {
   THD *thd= join->thd;
   COND_EQUAL *cond_equal= 0;
 
   if (cond) 
   {
-    cond= build_equal_items_for_cond(thd, cond, inherited);
+    cond= build_equal_items_for_cond(thd, cond, inherited, link_equal_fields);
     cond->update_used_tables();
     if (cond->type() == Item::COND_ITEM &&
         ((Item_cond*) cond)->functype() == Item_func::COND_AND_FUNC)
@@ -13548,9 +13819,10 @@ void optimize_wo_join_buffering(JOIN *join, uint first_tab, uint last_tab,
 
 
 static COND *
-optimize_cond(JOIN *join, COND *conds, 
+optimize_cond(JOIN *join, COND *conds,
               List<TABLE_LIST> *join_list, bool ignore_on_conds,
-              Item::cond_result *cond_value, COND_EQUAL **cond_equal)
+              Item::cond_result *cond_value, COND_EQUAL **cond_equal,
+              int flags)
 {
   THD *thd= join->thd;
   DBUG_ENTER("optimize_cond");
@@ -13573,9 +13845,10 @@ optimize_cond(JOIN *join, COND *conds,
       multiple equality contains a constant.
     */ 
     DBUG_EXECUTE("where", print_where(conds, "original", QT_ORDINARY););
-    conds= build_equal_items(join, conds, NULL, join_list, ignore_on_conds,
-                             cond_equal);
-     DBUG_EXECUTE("where",print_where(conds,"after equal_items", QT_ORDINARY););
+    conds= build_equal_items(join, conds, NULL, join_list, 
+                             ignore_on_conds, cond_equal,
+                             test(flags & OPT_LINK_EQUAL_FIELDS));
+    DBUG_EXECUTE("where",print_where(conds,"after equal_items", QT_ORDINARY););
 
     /* change field = field to field = const for each found field = const */
     propagate_cond_constants(thd, (I_List<COND_CMP> *) 0, conds, conds);
@@ -14068,6 +14341,8 @@ Field *create_tmp_field_from_field(THD *thd, Field *org_field,
       ((Field_double *) new_field)->not_fixed= TRUE;
     new_field->vcol_info= 0;
     new_field->stored_in_db= TRUE;
+    new_field->cond_selectivity= 1.0;
+    new_field->next_equal_field= NULL;
   }
   return new_field;
 }
@@ -14412,6 +14687,9 @@ void setup_tmp_table_column_bitmaps(TABLE *table, uchar *bitmaps)
   bitmap_init(&table->eq_join_set,
               (my_bitmap_map*) (bitmaps+ 3*bitmap_buffer_size(field_count)),
               field_count, FALSE);
+  bitmap_init(&table->cond_set,
+              (my_bitmap_map*) (bitmaps+ 4*bitmap_buffer_size(field_count)),
+              field_count, FALSE);
   /* write_set and all_set are copies of read_set */
   table->def_write_set= table->def_read_set;
   table->s->all_set= table->def_read_set;
@@ -14575,7 +14853,7 @@ create_tmp_table(THD *thd, TMP_TABLE_PARAM *param, List<Item> &fields,
                         &tmpname, (uint) strlen(path)+1,
                         &group_buff, (group && ! using_unique_constraint ?
                                       param->group_length : 0),
-                        &bitmaps, bitmap_buffer_size(field_count)*4,
+                        &bitmaps, bitmap_buffer_size(field_count)*5,
                         NullS))
   {
     if (temp_pool_slot != MY_BIT_NONE)
@@ -15330,7 +15608,7 @@ TABLE *create_virtual_tmp_table(THD *thd, List<Create_field> &field_list)
                         &share, sizeof(*share),
                         &field, (field_count + 1) * sizeof(Field*),
                         &blob_field, (field_count+1) *sizeof(uint),
-                        &bitmaps, bitmap_buffer_size(field_count)*4,
+                        &bitmaps, bitmap_buffer_size(field_count)*5,
                         NullS))
     return 0;
 
@@ -22221,7 +22499,13 @@ int JOIN::print_explain(select_result_sink *result, uint8 explain_flags,
         {
           float f= 0.0; 
           if (examined_rows)
-            f= (float) (100.0 * tab->records_read / examined_rows);
+	  {
+            double pushdown_cond_selectivity= tab->cond_selectivity;	      
+            if (pushdown_cond_selectivity == 1.0)
+              f= (float) (100.0 * tab->records_read / examined_rows);
+            else
+              f= (float) (100.0 * pushdown_cond_selectivity);
+          }
  	  set_if_smaller(f, 100.0);
           item_list.push_back(new Item_float(f, 2));
         }

sql/sql_select.h

@@ -292,6 +292,9 @@ typedef struct st_join_table {
   /* psergey-todo: make the below have type double, like POSITION::records_read? */
   ha_rows       records_read;
   
+  /* The selectivity of the conditions that can be pushed to the table */ 
+  double        cond_selectivity;  
+  
   /* Startup cost for execution */
   double        startup_cost;
     
@@ -772,6 +775,9 @@ typedef struct st_position :public Sql_alloc
   */
   double records_read;
 
+  /* The selectivity of the pushed down conditions */
+  double cond_selectivity; 
+
   /* 
     Cost accessing the table in course of the entire complete join execution,
     i.e. cost of one access method use (e.g. 'range' or 'ref' scan ) times 
@@ -1820,6 +1826,8 @@ void eliminate_tables(JOIN *join);
 /* Index Condition Pushdown entry point function */
 void push_index_cond(JOIN_TAB *tab, uint keyno);
 
+#define OPT_LINK_EQUAL_FIELDS    1
+
 /****************************************************************************
   Temporary table support for SQL Runtime
  ***************************************************************************/

sql/sql_statistics.cc

@@ -26,6 +26,7 @@
 #include "sql_base.h"
 #include "key.h"
 #include "sql_statistics.h"
+#include "opt_range.h"
 #include "my_atomic.h"
 
 /*
@@ -249,12 +250,13 @@ public:
   and column_name with column_name taken out of the only parameter f of the
   Field* type passed to this method. After this get_stat_values looks
   for a row by the set key value. If the row is found the values of statistical 
-  data columns min_value, max_value, nulls_ratio, avg_length, avg_frequency
-  are read into internal structures. Values of nulls_ratio, avg_length,
-  avg_frequency are read into the corresponding fields of the read_stat
-  structure from the Field object f, while values from min_value and max_value
-  are copied into the min_value and  max_value record buffers attached to the
-  TABLE structure for table t.
+  data columns min_value, max_value, nulls_ratio, avg_length, avg_frequency,
+  hist_size, hist_type, histogram are read into internal structures. Values
+  of nulls_ratio, avg_length, avg_frequency, hist_size, hist_type, histogram
+  are read into the corresponding fields of the read_stat  structure from
+  the Field object f, while values from min_value and max_value  are copied
+  into the min_value and  max_value record buffers attached to the TABLE
+   structure for table t.
   If the value of a statistical column in the found row is null, then the
   corresponding flag in the f->read_stat.column_stat_nulls bitmap is set off.
   Otherwise the flag is set on. If no row is found for the column the all flags
@@ -867,11 +869,12 @@ public:
 
     @details
     This implementation of a purely virtual method sets the value of the
-    columns 'min_value', 'max_value', 'nulls_ratio', 'avg_length' and
-    'avg_frequency' of the stistical table columns_stat according to the
-    contents of the bitmap write_stat.column_stat_nulls and the values
-    of the fields min_value, max_value, nulls_ratio, avg_length and
-    avg_frequency of the structure write_stat from the Field structure
+    columns 'min_value', 'max_value', 'nulls_ratio', 'avg_length',
+    'avg_frequency', 'hist_size', 'hist_type' and 'histogram'  of the 
+    stistical table columns_stat according to the contents of the bitmap
+    write_stat.column_stat_nulls and the values of the fields min_value,
+    max_value, nulls_ratio, avg_length, avg_frequency, hist_size, hist_type
+    and histogram of the structure write_stat from the Field structure
     for the field 'table_field'.
     The value of the k-th column in the table columns_stat is set to NULL
     if the k-th bit in the bitmap 'column_stat_nulls' is set to 1. 
@@ -888,7 +891,7 @@ public:
     char buff[MAX_FIELD_WIDTH];
     String val(buff, sizeof(buff), &my_charset_utf8_bin);
 
-    for (uint i= COLUMN_STAT_MIN_VALUE; i <= COLUMN_STAT_AVG_FREQUENCY; i++)
+    for (uint i= COLUMN_STAT_MIN_VALUE; i <= COLUMN_STAT_HISTOGRAM; i++)
     {  
       Field *stat_field= stat_table->field[i];
       if (table_field->collected_stats->is_null(i))
@@ -923,7 +926,21 @@ public:
           break;
         case COLUMN_STAT_AVG_FREQUENCY:
           stat_field->store(table_field->collected_stats->get_avg_frequency());
-          break;            
+          break; 
+        case COLUMN_STAT_HIST_SIZE:
+          stat_field->store(table_field->collected_stats->histogram.get_size());
+          break;
+        case COLUMN_STAT_HIST_TYPE:
+          stat_field->store(table_field->collected_stats->histogram.get_type() +
+                            1);
+          break;
+        case COLUMN_STAT_HISTOGRAM:
+          const char * col_histogram=
+          (const char *) (table_field->collected_stats->histogram.get_values());
+	  stat_field->store(col_histogram,
+                            table_field->collected_stats->histogram.get_size(),
+                            &my_charset_bin);
+          break;           
         }
       }
     }
@@ -936,14 +953,15 @@ public:
 
     @details
     This implementation of a purely virtual method first looks for a record
-    the statistical table column_stats by its primary key set the record
+    in the statistical table column_stats by its primary key set in the record
     buffer with the help of Column_stat::set_key_fields. Then, if the row is
     found, the function reads the values of the columns 'min_value',
-    'max_value', 'nulls_ratio', 'avg_length' and 'avg_frequency' of the
-    table column_stat and sets accordingly the value of the bitmap 
-    read_stat.column_stat_nulls' and the values of the fields min_value,
-    max_value, nulls_ratio, avg_length and avg_frequency of the structure
-    read_stat from the Field structure for the field 'table_field'.
+    'max_value', 'nulls_ratio', 'avg_length', 'avg_frequency', 'hist_size' and
+    'hist_type" of the  table column_stat and sets accordingly the value of
+    the bitmap  read_stat.column_stat_nulls' and the values of the fields
+    min_value, max_value, nulls_ratio, avg_length, avg_frequency, hist_size and
+    hist_type of the structure read_stat from the Field structure for the field
+    'table_field'.
   */    
 
   void get_stat_values()
@@ -960,7 +978,7 @@ public:
       char buff[MAX_FIELD_WIDTH];
       String val(buff, sizeof(buff), &my_charset_utf8_bin);
 
-      for (uint i= COLUMN_STAT_MIN_VALUE; i <= COLUMN_STAT_AVG_FREQUENCY; i++)
+      for (uint i= COLUMN_STAT_MIN_VALUE; i <= COLUMN_STAT_HIST_TYPE; i++)
       {  
         Field *stat_field= stat_table->field[i];
 
@@ -992,6 +1010,14 @@ public:
             break;
           case COLUMN_STAT_AVG_FREQUENCY:
             table_field->read_stats->set_avg_frequency(stat_field->val_real());
+            break;
+          case COLUMN_STAT_HIST_SIZE:
+            table_field->read_stats->histogram.set_size(stat_field->val_int());
+            break;            
+          case COLUMN_STAT_HIST_TYPE:
+            Histogram_type hist_type= (Histogram_type) (stat_field->val_int() -
+                                                        1);
+            table_field->read_stats->histogram.set_type(hist_type);
             break;            
           }
         }
@@ -999,6 +1025,37 @@ public:
     }
   }
 
+
+  /** 
+    @brief
+    Read histogram from of column_stats
+
+    @details
+    This method first looks for a record in the statistical table column_stats
+    by its primary key set the record buffer with the help of
+    Column_stat::set_key_fields. Then, if the row is found, the function reads
+    the value of the column 'histogram' of the  table column_stat and sets
+    accordingly the corresponding bit in the bitmap read_stat.column_stat_nulls.
+    The method assumes that the value of histogram size and the pointer to
+    the histogram location has been already set in the fields size and values
+    of read_stats->histogram.
+  */    
+
+  void get_histogram_value()
+  {
+    if (find_stat())
+    {
+      char buff[MAX_FIELD_WIDTH];
+      String val(buff, sizeof(buff), &my_charset_utf8_bin);
+      uint fldno= COLUMN_STAT_HISTOGRAM;
+      Field *stat_field= stat_table->field[fldno];
+      table_field->read_stats->set_not_null(fldno);
+      stat_field->val_str(&val);
+      memcpy(table_field->read_stats->histogram.get_values(),
+             val.ptr(), table_field->read_stats->histogram.get_size());
+    }
+  }
+
 };
 
 
@@ -1200,6 +1257,76 @@ public:
 
 };
 
+/*
+  Histogram_builder is a helper class that is used to build histograms
+  for columns
+*/
+
+class Histogram_builder
+{
+  Field *column;           /* table field for which the histogram is built */
+  uint col_length;         /* size of this field                           */
+  ha_rows records;         /* number of records the histogram is built for */
+  Field *min_value;        /* pointer to the minimal value for the field   */
+  Field *max_value;        /* pointer to the maximal value for the field   */
+  Histogram *histogram;    /* the histogram location                       */
+  uint hist_width;         /* the number of points in the histogram        */
+  double bucket_capacity;  /* number of rows in a bucket of the histogram  */ 
+  uint curr_bucket;        /* number of the current bucket to be built     */
+  ulonglong count;         /* number of values retrieved                   */
+  ulonglong count_distinct;    /* number of distinct values retrieved      */
+
+public: 
+  Histogram_builder(Field *col, uint col_len, ha_rows rows)
+    : column(col), col_length(col_len), records(rows)
+  {
+    Column_statistics *col_stats= col->collected_stats;
+    min_value= col_stats->min_value;
+    max_value= col_stats->max_value;
+    histogram= &col_stats->histogram;
+    hist_width= histogram->get_width();
+    bucket_capacity= (double) records / (hist_width + 1);
+    curr_bucket= 0;
+    count= 0;
+    count_distinct= 0;    
+  }
+
+  ulonglong get_count_distinct() { return count_distinct; }
+
+  int next(void *elem, element_count elem_cnt)
+  {
+    count_distinct++;
+    count+= elem_cnt;
+    if (curr_bucket == hist_width)
+      return 0;
+    if (count > bucket_capacity * (curr_bucket + 1))
+    {
+      column->store_field_value((uchar *) elem, col_length);
+      histogram->set_value(curr_bucket,
+                           column->pos_in_interval(min_value, max_value)); 
+      curr_bucket++;
+      while (curr_bucket != hist_width &&
+             count > bucket_capacity * (curr_bucket + 1))
+      {
+        histogram->set_prev_value(curr_bucket);
+	curr_bucket++;
+      }
+    }
+    return 0;
+  }
+};
+
+
+C_MODE_START
+
+int histogram_build_walk(void *elem, element_count elem_cnt, void *arg)
+{
+  Histogram_builder *hist_builder= (Histogram_builder *) arg;
+  return hist_builder->next(elem, elem_cnt);
+}
+
+C_MODE_END
+
 
 /*
   The class Count_distinct_field is a helper class used to calculate
@@ -1220,6 +1347,8 @@ protected:
   uint tree_key_length; /* The length of the keys for the elements of 'tree */
 
 public:
+  
+  Count_distinct_field() {}
 
   /**
     @param
@@ -1238,28 +1367,11 @@ public:
 
   Count_distinct_field(Field *field, uint max_heap_table_size)
   {
-    qsort_cmp2 compare_key;
-    void* cmp_arg;
-    enum enum_field_types f_type= field->type();
-
     table_field= field;
     tree_key_length= field->pack_length();
 
-    if ((f_type == MYSQL_TYPE_VARCHAR) ||
-	(!field->binary() && (f_type == MYSQL_TYPE_STRING ||
-                              f_type == MYSQL_TYPE_VAR_STRING)))
-    {
-      compare_key= (qsort_cmp2) simple_str_key_cmp;
-      cmp_arg= (void*) field;      
-    }
-    else
-    {
-      cmp_arg= (void*) &tree_key_length;
-      compare_key= (qsort_cmp2) simple_raw_key_cmp;
-    }
-        
-    tree= new Unique(compare_key, cmp_arg,
-                     tree_key_length, max_heap_table_size);
+    tree= new Unique((qsort_cmp2) simple_str_key_cmp, (void*) field,
+                     tree_key_length, max_heap_table_size, 1);
   }
 
   virtual ~Count_distinct_field()
@@ -1299,9 +1411,48 @@ public:
     tree->walk(table_field->table, count_distinct_walk, (void*) &count);
     return count;
   }
+
+  /*
+    @brief
+    Build the histogram for the elements accumulated in the container of 'tree'
+  */
+  ulonglong get_value_with_histogram(ha_rows rows)
+  {
+    Histogram_builder hist_builder(table_field, tree_key_length, rows);
+    tree->walk(table_field->table,  histogram_build_walk, (void *) &hist_builder);
+    return hist_builder.get_count_distinct();
+  }
+
+  /*
+    @brief
+    Get the size of the histogram in bytes built for table_field
+  */
+  uint get_hist_size()
+  {
+    return table_field->collected_stats->histogram.get_size();
+  }
+
+  /*
+    @brief
+    Get the pointer to the histogram built for table_field
+  */
+  uchar *get_histogram()
+  {
+    return table_field->collected_stats->histogram.get_values();
+  }
+
 };
 
 
+static
+int simple_ulonglong_key_cmp(void* arg, uchar* key1, uchar* key2)
+{
+  ulonglong *val1= (ulonglong *) key1;
+  ulonglong *val2= (ulonglong *) key2;
+  return *val1 > *val2 ? 1 : *val1 == *val2 ? 0 : -1; 
+}
+  
+
 /* 
   The class Count_distinct_field_bit is derived from the class 
   Count_distinct_field to be used only for fields of the MYSQL_TYPE_BIT type.
@@ -1311,8 +1462,17 @@ public:
 class Count_distinct_field_bit: public Count_distinct_field
 {
 public:
+
   Count_distinct_field_bit(Field *field, uint max_heap_table_size)
-    :Count_distinct_field(field, max_heap_table_size) {}
+  {
+    table_field= field;
+    tree_key_length= sizeof(ulonglong);
+
+    tree= new Unique((qsort_cmp2) simple_ulonglong_key_cmp,
+                     (void*) &tree_key_length,
+                     tree_key_length, max_heap_table_size, 1);
+  }
+
   bool add()
   {
     longlong val= table_field->val_int();   
@@ -1671,13 +1831,27 @@ int alloc_statistics_for_table(THD* thd, TABLE *table)
   ulong *idx_avg_frequency= (ulong*) alloc_root(&table->mem_root,
                                                 sizeof(ulong) * key_parts);
 
-  if (!table_stats || !column_stats || !index_stats || !idx_avg_frequency)
+  uint columns= 0;
+  for (field_ptr= table->field; *field_ptr; field_ptr++)
+  {
+    if (bitmap_is_set(table->read_set, (*field_ptr)->field_index))
+      columns++;
+  }
+  uint hist_size= thd->variables.histogram_size;
+  Histogram_type hist_type= (Histogram_type) (thd->variables.histogram_type);
+  uchar *histogram= NULL;
+  if (hist_size > 0)
+    histogram= (uchar *) alloc_root(&table->mem_root, hist_size * columns);
+
+  if (!table_stats || !column_stats || !index_stats || !idx_avg_frequency ||
+      (hist_size && !histogram))
     DBUG_RETURN(1);
 
   table->collected_stats= table_stats;
   table_stats->column_stats= column_stats;
   table_stats->index_stats= index_stats;
   table_stats->idx_avg_frequency= idx_avg_frequency;
+  table_stats->histograms= histogram;
   
   memset(column_stats, 0, sizeof(Column_statistics) * (fields+1));
 
@@ -1686,6 +1860,13 @@ int alloc_statistics_for_table(THD* thd, TABLE *table)
     (*field_ptr)->collected_stats= column_stats;
     (*field_ptr)->collected_stats->max_value= NULL;
     (*field_ptr)->collected_stats->min_value= NULL;
+    if (bitmap_is_set(table->read_set, (*field_ptr)->field_index))
+    {
+      column_stats->histogram.set_size(hist_size);
+      column_stats->histogram.set_type(hist_type);
+      column_stats->histogram.set_values(histogram);
+      histogram+= hist_size;
+    }
   }
 
   memset(idx_avg_frequency, 0, sizeof(ulong) * key_parts);
@@ -1790,7 +1971,6 @@ inline bool statistics_for_command_is_needed(THD *thd)
   @note
   Currently the function always is called with the parameter is_safe set
   to FALSE. 
-
 */      
 
 int alloc_statistics_for_table_share(THD* thd, TABLE_SHARE *table_share, 
@@ -1902,11 +2082,82 @@ int alloc_statistics_for_table_share(THD* thd, TABLE_SHARE *table_share,
   if (!is_safe)
     mysql_mutex_unlock(&table_share->LOCK_ha_data);
 
-
   DBUG_RETURN(0);
 }
 
 
+/**
+  @brief 
+  Allocate memory for the histogram used by a table share
+
+  @param
+  thd         Thread handler
+  @param
+  table_share Table share for which the memory for histogram data is allocated
+  @param
+  is_safe     TRUE <-> at any time only one thread can perform the function
+
+  @note
+  The function allocates the memory for the histogram built for a table in the
+  table's share memory with the intention to read the data there from the
+  system persistent statistical table mysql.column_stats,
+  The memory is allocated in the table_share's mem_root.
+  If the parameter is_safe is TRUE then it is guaranteed that at any given time
+  only one thread is executed the code of the function.
+
+  @retval
+  0     If the memory for all statistical data has been successfully allocated  
+  @retval
+  1     Otherwise
+
+  @note
+  Currently the function always is called with the parameter is_safe set
+  to FALSE. 
+*/      
+
+static
+int alloc_histograms_for_table_share(THD* thd, TABLE_SHARE *table_share, 
+                                     bool is_safe)
+{
+  TABLE_STATISTICS_CB *stats_cb= &table_share->stats_cb;
+
+  DBUG_ENTER("alloc_histograms_for_table_share");
+
+  if (!is_safe)
+    mysql_mutex_lock(&table_share->LOCK_ha_data);
+
+  if (stats_cb->histograms_can_be_read)
+  {
+    if (!is_safe)
+      mysql_mutex_unlock(&table_share->LOCK_ha_data);
+    DBUG_RETURN(0);
+  }
+
+  Table_statistics *table_stats= stats_cb->table_stats;
+  ulong total_hist_size= table_stats->total_hist_size;
+
+  if (total_hist_size && !table_stats->histograms)
+  {
+    uchar *histograms= (uchar *) alloc_root(&stats_cb->mem_root,
+                                            total_hist_size);
+    if (!histograms)
+    {
+      if (!is_safe)
+        mysql_mutex_unlock(&table_share->LOCK_ha_data);
+      DBUG_RETURN(1);
+    }
+    memset(histograms, 0, total_hist_size);
+    table_stats->histograms= histograms;
+    stats_cb->histograms_can_be_read= TRUE;
+  }
+
+  if (!is_safe)
+    mysql_mutex_unlock(&table_share->LOCK_ha_data);
+
+  DBUG_RETURN(0);
+
+}
+
 /**
   @brief
   Initialize the aggregation fields to collect statistics on a column
@@ -2005,14 +2256,29 @@ void Column_statistics_collected::finish(ha_rows rows)
      set_not_null(COLUMN_STAT_AVG_LENGTH);
   }
   if (count_distinct)
-  { 
-    ulonglong distincts= count_distinct->get_value();
+  {
+    ulonglong distincts;
+    uint hist_size= count_distinct->get_hist_size();
+    if (hist_size == 0)
+      distincts= count_distinct->get_value();
+    else
+      distincts= count_distinct->get_value_with_histogram(rows - nulls);
     if (distincts)
     {
       val= (double) (rows - nulls) / distincts;
       set_avg_frequency(val); 
       set_not_null(COLUMN_STAT_AVG_FREQUENCY);
     }
+    else
+      hist_size= 0;
+    histogram.set_size(hist_size);
+    set_not_null(COLUMN_STAT_HIST_SIZE);
+    if (hist_size && distincts)
+    {
+      set_not_null(COLUMN_STAT_HIST_TYPE);
+      histogram.set_values(count_distinct->get_histogram());
+      set_not_null(COLUMN_STAT_HISTOGRAM);
+    } 
     delete count_distinct;
     count_distinct= NULL;
   }
@@ -2233,16 +2499,19 @@ int collect_statistics_for_table(THD *thd, TABLE *table)
     table->collected_stats->cardinality= rows;
   }
 
+  bitmap_clear_all(table->write_set);
   for (field_ptr= table->field; *field_ptr; field_ptr++)
   {
     table_field= *field_ptr;
     if (!bitmap_is_set(table->read_set, table_field->field_index))
       continue;
+    bitmap_set_bit(table->write_set, table_field->field_index); 
     if (!rc)
       table_field->collected_stats->finish(rows);
     else
       table_field->collected_stats->cleanup();
   }
+bitmap_clear_all(table->write_set);
 
   if (!rc)
   {
@@ -2420,6 +2689,7 @@ int read_statistics_for_table(THD *thd, TABLE *table, TABLE_LIST *stat_tables)
   Field **field_ptr;
   KEY *key_info, *key_info_end;
   TABLE_SHARE *table_share= table->s;
+  Table_statistics *read_stats= table_share->stats_cb.table_stats;
 
   DBUG_ENTER("read_statistics_for_table");
 
@@ -2431,16 +2701,18 @@ int read_statistics_for_table(THD *thd, TABLE *table, TABLE_LIST *stat_tables)
    
   /* Read statistics from the statistical table column_stats */
   stat_table= stat_tables[COLUMN_STAT].table;
+  ulong total_hist_size= 0;
   Column_stat column_stat(stat_table, table);
   for (field_ptr= table_share->field; *field_ptr; field_ptr++)
   {
     table_field= *field_ptr;
     column_stat.set_key_fields(table_field);
     column_stat.get_stat_values();
+    total_hist_size+= table_field->read_stats->histogram.get_size();
   }
+  read_stats->total_hist_size= total_hist_size;
 
   /* Read statistics from the statistical table index_stats */
-  Table_statistics *read_stats= table_share->stats_cb.table_stats;
   stat_table= stat_tables[INDEX_STAT].table;
   Index_stat index_stat(stat_table, table);
   for (key_info= table_share->key_info,
@@ -2558,10 +2830,14 @@ bool statistics_for_tables_is_needed(THD *thd, TABLE_LIST *tables)
       TABLE_SHARE *table_share= tl->table->s;
       if (table_share && 
           table_share->stats_cb.stats_can_be_read &&
-          !table_share->stats_cb.stats_is_read)
+          (!table_share->stats_cb.stats_is_read ||
+           (!table_share->stats_cb.histograms_are_read &&
+            thd->variables.optimizer_use_condition_selectivity > 3)))
         return TRUE;
       if (table_share->stats_cb.stats_is_read)
         tl->table->stats_is_read= TRUE;
+      if (table_share->stats_cb.histograms_are_read)
+        tl->table->histograms_are_read= TRUE;
     } 
   }
 
@@ -2569,6 +2845,73 @@ bool statistics_for_tables_is_needed(THD *thd, TABLE_LIST *tables)
 }
 
 
+/**
+  @brief
+  Read histogram for a table from the persistent statistical tables
+
+  @param
+  thd         The thread handle
+  @param
+  table       The table to read histograms for
+  @param
+  stat_tables The array of TABLE_LIST objects for statistical tables
+
+  @details
+  For the statistical table columns_stats the function looks for the rows
+  from this table that contain statistical data on 'table'. If such rows
+  are found the histograms from them are read into the memory allocated
+  for histograms of 'table'. Later at the query processing these histogram
+  are supposed to be used by the optimizer. 
+  The parameter stat_tables should point to an array of TABLE_LIST
+  objects for all statistical tables linked into a list. All statistical
+  tables are supposed to be opened.  
+  The function is called by read_statistics_for_tables_if_needed().
+
+  @retval
+  0         If data has been successfully read for the table  
+  @retval
+  1         Otherwise
+
+  @note
+  Objects of the helper Column_stat are employed read histogram
+  from the statistical table column_stats now.        
+*/
+
+static
+int read_histograms_for_table(THD *thd, TABLE *table, TABLE_LIST *stat_tables)
+{
+  TABLE_SHARE *table_share= table->s;
+  
+  DBUG_ENTER("read_histograms_for_table");
+
+  if (!table_share->stats_cb.histograms_can_be_read)
+  {
+    (void) alloc_histograms_for_table_share(thd, table_share, FALSE);
+  }
+  if (table_share->stats_cb.histograms_can_be_read &&
+      !table_share->stats_cb.histograms_are_read)
+  {
+    Field **field_ptr;
+    uchar *histogram= table_share->stats_cb.table_stats->histograms;
+    TABLE *stat_table= stat_tables[COLUMN_STAT].table;
+    Column_stat column_stat(stat_table, table);
+    for (field_ptr= table_share->field; *field_ptr; field_ptr++)
+    {
+      Field *table_field= *field_ptr;
+      uint hist_size= table_field->read_stats->histogram.get_size();
+      if (hist_size)
+      {
+        column_stat.set_key_fields(table_field);
+        table_field->read_stats->histogram.set_values(histogram);
+        column_stat.get_histogram_value();
+        histogram+= hist_size;
+      }
+    }
+  }
+ 
+  DBUG_RETURN(0);
+}
+
 /**
   @brief
   Read statistics for tables from a table list if it is needed
@@ -2596,7 +2939,7 @@ int read_statistics_for_tables_if_needed(THD *thd, TABLE_LIST *tables)
   TABLE_LIST stat_tables[STATISTICS_TABLES];
   Open_tables_backup open_tables_backup;
 
-  DBUG_ENTER("read_statistics_for_table_if_needed");
+  DBUG_ENTER("read_statistics_for_tables_if_needed");
 
   DEBUG_SYNC(thd, "statistics_read_start");
 
@@ -2623,6 +2966,14 @@ int read_statistics_for_tables_if_needed(THD *thd, TABLE_LIST *tables)
       }
       if (table_share->stats_cb.stats_is_read)
         tl->table->stats_is_read= TRUE;
+      if (thd->variables.optimizer_use_condition_selectivity > 3 && 
+          table_share && !table_share->stats_cb.histograms_are_read)
+      {
+        (void) read_histograms_for_table(thd, tl->table, stat_tables);
+        table_share->stats_cb.histograms_are_read= TRUE;
+      }
+      if (table_share->stats_cb.stats_is_read)
+        tl->table->histograms_are_read= TRUE;
     }
   }  
 
@@ -3054,3 +3405,152 @@ void set_statistics_for_table(THD *thd, TABLE *table)
        key_info->read_stats->get_avg_frequency(0) > 0.5);
   }
 }
+
+
+/**
+  @brief
+  Get the average frequency for a column 
+
+  @param
+  field       The column whose average frequency is required
+
+  @retval
+  The required average frequency
+*/
+
+double get_column_avg_frequency(Field * field)
+{
+  double res;
+  TABLE *table= field->table;
+
+  /* 
+    Statistics is shared by table instances and  is accessed through
+    the table share. If table->s->field is not set for 'table', then
+    no column statistics is available for the table .
+  */
+  if (!table->s->field)
+  {
+    res= table->stat_records();
+    return res;
+  }
+ 
+  Column_statistics *col_stats= table->s->field[field->field_index]->read_stats;
+
+  if (!col_stats)
+    res= table->stat_records();
+  else
+    res= col_stats->get_avg_frequency();
+  return res;
+} 
+
+
+/**
+  @brief
+  Estimate the number of rows in a column range using data from stat tables 
+
+  @param
+  field       The column whose range cardinality is to be estimated
+  @param
+  min_endp    The left end of the range whose cardinality is required 
+  @param
+  max_endp    The right end of the range whose cardinality is required 
+  @param
+  range_flag  The range flags
+
+  @details
+  The function gets an estimate of the number of rows in a column range
+  using the statistical data from the table column_stats.
+
+  @retval
+  The required estimate of the rows in the column range
+*/
+
+double get_column_range_cardinality(Field *field,
+                                    key_range *min_endp,
+                                    key_range *max_endp,
+                                    uint range_flag)
+{
+  double res;
+  TABLE *table= field->table;
+  Column_statistics *col_stats= table->field[field->field_index]->read_stats;
+  double tab_records= table->stat_records();
+
+  if (!col_stats)
+    return tab_records;
+
+  double col_nulls= tab_records * col_stats->get_nulls_ratio();
+
+  double col_non_nulls= tab_records - col_nulls;
+
+  bool nulls_incl= field->null_ptr && min_endp && min_endp->key[0] &&
+                   !(range_flag & NEAR_MIN);
+
+  if (col_non_nulls < 1)
+    res= 0;
+  else if (min_endp && max_endp && min_endp->length == max_endp->length &&
+           !memcmp(min_endp->key, max_endp->key, min_endp->length))
+  { 
+    if (nulls_incl)
+    {
+      /* This is null single point range */
+      res= col_nulls;
+    }
+    else
+    {
+      double avg_frequency= col_stats->get_avg_frequency();
+      res= avg_frequency;   
+      if (avg_frequency > 1.0 + 0.000001 && 
+          col_stats->min_value && col_stats->max_value)
+      {
+        Histogram *hist= &col_stats->histogram;
+        if (hist->is_available())
+        {
+          double pos= field->pos_in_interval(col_stats->min_value,
+                                             col_stats->max_value);
+          res= col_non_nulls * 
+	       hist->point_selectivity(pos,
+                                       avg_frequency / col_non_nulls);
+        }
+      }
+    }
+  }  
+  else 
+  {
+    if (col_stats->min_value && col_stats->max_value)
+    {
+      double sel, min_mp_pos, max_mp_pos;
+
+      if (min_endp && !min_endp->key[0])
+      {
+        store_key_image_to_rec(field, (uchar *) min_endp->key,
+                               min_endp->length);
+        min_mp_pos= field->pos_in_interval(col_stats->min_value,
+                                           col_stats->max_value);
+      }
+      else
+        min_mp_pos= 0.0;
+      if (max_endp)
+      {
+        store_key_image_to_rec(field, (uchar *) max_endp->key,
+                               max_endp->length);
+        max_mp_pos= field->pos_in_interval(col_stats->min_value,
+                                           col_stats->max_value);
+      }
+      else
+        max_mp_pos= 1.0;
+
+      Histogram *hist= &col_stats->histogram;
+      if (!hist->is_available())
+        sel= (max_mp_pos - min_mp_pos);
+      else
+        sel= hist->range_selectivity(min_mp_pos, max_mp_pos);
+      res= col_non_nulls * sel;
+      set_if_bigger(res, col_stats->get_avg_frequency());
+    }
+    else
+      res= col_non_nulls;
+    if (nulls_incl)
+      res+= col_nulls;
+  }
+  return res;
+}

sql/sql_statistics.h

@@ -16,15 +16,6 @@
 #ifndef SQL_STATISTICS_H
 #define SQL_STATISTICS_H
 
-/* 
-  These enumeration types comprise the dictionary of three
-  statistical tables table_stat, column_stat and index_stat
-  as they defined in ../scripts/mysql_system_tables.sql.
-
-  It would be nice if the declarations of these types were
-  generated automatically by the table definitions.   
-*/
-
 typedef
 enum enum_use_stat_tables_mode
 {
@@ -33,6 +24,13 @@ enum enum_use_stat_tables_mode
   PEFERABLY,
 } Use_stat_tables_mode;
 
+typedef
+enum enum_histogram_type
+{
+  SINGLE_PREC_HB,
+  DOUBLE_PREC_HB
+} Histogram_type;
+
 enum enum_stat_tables
 {
   TABLE_STAT,
@@ -40,6 +38,16 @@ enum enum_stat_tables
   INDEX_STAT,
 };
 
+
+/* 
+  These enumeration types comprise the dictionary of three
+  statistical tables table_stat, column_stat and index_stat
+  as they defined in ../scripts/mysql_system_tables.sql.
+
+  It would be nice if the declarations of these types were
+  generated automatically by the table definitions.   
+*/
+
 enum enum_table_stat_col
 {
   TABLE_STAT_DB_NAME,
@@ -56,7 +64,10 @@ enum enum_column_stat_col
   COLUMN_STAT_MAX_VALUE,
   COLUMN_STAT_NULLS_RATIO,
   COLUMN_STAT_AVG_LENGTH,
-  COLUMN_STAT_AVG_FREQUENCY
+  COLUMN_STAT_AVG_FREQUENCY,
+  COLUMN_STAT_HIST_SIZE,
+  COLUMN_STAT_HIST_TYPE,
+  COLUMN_STAT_HISTOGRAM
 };
 
 enum enum_index_stat_col
@@ -90,6 +101,160 @@ int rename_column_in_stat_tables(THD *thd, TABLE *tab, Field *col,
                                   const char *new_name);
 void set_statistics_for_table(THD *thd, TABLE *table);
 
+double get_column_avg_frequency(Field * field);
+
+double get_column_range_cardinality(Field *field,
+                                    key_range *min_endp,
+                                    key_range *max_endp,
+                                    uint range_flag);
+
+class Histogram
+{
+
+private:
+  Histogram_type type;
+  uint8 size;
+  uchar *values;
+
+  uint prec_factor()
+  {
+    switch (type) {
+    case SINGLE_PREC_HB:
+      return ((uint) (1 << 8) - 1);
+    case DOUBLE_PREC_HB:
+      return ((uint) (1 << 16) - 1);
+    }
+    return 1;
+  }
+
+public:
+  uint get_width()
+  {
+    switch (type) {
+    case SINGLE_PREC_HB:
+      return size;
+    case DOUBLE_PREC_HB:
+      return size / 2;
+    }
+    return 0;
+  }
+
+private:
+  uint get_value(uint i)
+  {
+    switch (type) {
+    case SINGLE_PREC_HB:
+      return (uint) (((uint8 *) values)[i]);
+    case DOUBLE_PREC_HB:
+      return (uint) (((uint16 *) values)[i]);
+    }
+    return 0;
+  }
+
+  uint find_bucket(double pos, bool first)
+  {
+    uint val= (uint) (pos * prec_factor());
+    int lp= 0;
+    int rp= get_width() - 1;
+    int d= get_width() / 2;
+    uint i= lp + d;
+    for ( ; d;  d= (rp - lp) / 2, i= lp + d)
+    {
+      if (val == get_value(i))
+	break; 
+      if (val < get_value(i))
+        rp= i;
+      else if (val > get_value(i + 1))
+        lp= i + 1;
+      else
+        break;
+    }
+    if (val == get_value(i))
+    {
+      if (first)
+      {
+        while(i && val == get_value(i - 1))
+          i--;
+      }
+      else
+      {
+        while(i + 1 < get_width() && val == get_value(i + 1))
+          i++;
+      }
+    }
+    return i;
+  }
+
+public:
+
+  uint get_size() { return (uint) size; }
+
+  Histogram_type get_type() { return type; }
+
+  uchar *get_values() { return (uchar *) values; }
+
+  void set_size (ulonglong sz) { size= (uint8) sz; }
+
+  void set_type (Histogram_type t) { type= t; }
+
+  void set_values (uchar *vals) { values= (uchar *) vals; }
+
+  bool is_available() { return get_size() > 0 && get_values(); }
+
+  void set_value(uint i, double val)
+  {
+    switch (type) {
+    case SINGLE_PREC_HB:   
+      ((uint8 *) values)[i]= (uint8) (val * prec_factor());
+      return;
+    case DOUBLE_PREC_HB:
+      ((uint16 *) values)[i]= (uint16) (val * prec_factor());
+      return;
+    }
+  }
+
+  void set_prev_value(uint i)
+  {
+    switch (type) {
+    case SINGLE_PREC_HB:   
+      ((uint8 *) values)[i]= ((uint8 *) values)[i-1];
+      return;
+    case DOUBLE_PREC_HB:
+      ((uint16 *) values)[i]= ((uint16 *) values)[i-1];
+      return;
+    }
+  }
+
+  double range_selectivity(double min_pos, double max_pos)
+  {
+    double sel;
+    double bucket_sel= 1.0/(get_width() + 1);  
+    uint min= find_bucket(min_pos, TRUE);
+    uint max= find_bucket(max_pos, FALSE);
+    sel= bucket_sel * (max - min + 1);
+    return sel;
+  } 
+
+  double point_selectivity(double pos, double avg_sel)
+  {
+    double sel;
+    double bucket_sel= 1.0/(get_width() + 1);  
+    uint min= find_bucket(pos, TRUE);
+    uint max= min;
+    while (max + 1 < get_width() && get_value(max + 1) == get_value(max))
+      max++;
+    double inv_prec_factor= (double) 1.0 / prec_factor(); 
+    double width= (max + 1 == get_width() ?
+                   1.0 : get_value(max) * inv_prec_factor) -
+	          (min == 0 ?
+                   0.0 : get_value(min-1) * inv_prec_factor); 
+    sel= avg_sel * (bucket_sel * (max + 1 - min)) / width;
+    return sel;
+  }
+             
+};
+
+
 class Columns_statistics;
 class Index_statistics;
 
@@ -105,8 +270,9 @@ public:
   uchar *min_max_record_buffers;    /* Record buffers for min/max values  */
   Column_statistics *column_stats;  /* Array of statistical data for columns */
   Index_statistics *index_stats;    /* Array of statistical data for indexes */
-  ulong *idx_avg_frequency;   /* Array of records per key for index prefixes */                                       
-
+  ulong *idx_avg_frequency;   /* Array of records per key for index prefixes */
+  ulong total_hist_size;            /* Total size of all histograms */
+  uchar *histograms;                /* Sequence of histograms       */                    
 };
 
 
@@ -161,10 +327,12 @@ private:
 
 public:
 
+  Histogram histogram;
+ 
   void set_all_nulls()
   {
     column_stat_nulls= 
-      ((1 << (COLUMN_STAT_AVG_FREQUENCY-COLUMN_STAT_COLUMN_NAME))-1) <<
+      ((1 << (COLUMN_STAT_HISTOGRAM-COLUMN_STAT_COLUMN_NAME))-1) <<
       (COLUMN_STAT_COLUMN_NAME+1);
   }
 

sql/sys_vars.cc

@@ -1533,6 +1533,25 @@ static Sys_var_ulong Sys_optimizer_prune_level(
        SESSION_VAR(optimizer_prune_level), CMD_LINE(REQUIRED_ARG),
        VALID_RANGE(0, 1), DEFAULT(1), BLOCK_SIZE(1));
 
+static Sys_var_ulong Sys_optimizer_use_condition_selectivity(
+       "optimizer_use_condition_selectivity",
+       "Controls selectivity of which conditions the optimizer takes into "
+       "account to calculate cardinality of a partial join when it searches "
+       "for the best execution plan "
+       "Meaning: "
+       "1 - use selectivity of index backed range conditions to calculate "
+       "the cardinality of a partial join if the last joined table is "
+       "accessed by full table scan or an index scan, "
+       "2 - use selectivity of index backed range conditions to calculate "
+       "the cardinality of a partial join in any case, "
+       "3 - additionally always use selectivity of range conditions that are "
+       "not backed by any index to calculate the cardinality of a partial join, "
+       "4 - use histograms to calculate selectivity of range conditions that "
+       "are not backed by any index to calculate the cardinality of "
+       "a partial join.",
+       SESSION_VAR(optimizer_use_condition_selectivity), CMD_LINE(REQUIRED_ARG),
+       VALID_RANGE(1, 4), DEFAULT(1), BLOCK_SIZE(1));
+
 /** Warns about deprecated value 63 */
 static bool fix_optimizer_search_depth(sys_var *self, THD *thd,
                                        enum_var_type type)
@@ -3876,6 +3895,24 @@ static Sys_var_enum Sys_optimizer_use_stat_tables(
        SESSION_VAR(use_stat_tables), CMD_LINE(REQUIRED_ARG),
        use_stat_tables_modes, DEFAULT(0));
 
+static Sys_var_ulong Sys_histogram_size(
+       "histogram_size",
+       "Number of bytes used for a histogram. "
+       "If set to 0, no histograms are created by ANALYZE.",
+       SESSION_VAR(histogram_size), CMD_LINE(REQUIRED_ARG),
+       VALID_RANGE(0, 255), DEFAULT(0), BLOCK_SIZE(1));
+
+const char *histogram_types[] =
+           {"SINGLE_PREC_HB", "DOUBLE_PREC_HB", 0};
+static Sys_var_enum Sys_histogram_type(
+       "histogram_type",
+       "Specifies type of the histograms created by ANALYZE. "
+       "Possible values are: "
+       "SINGLE_PREC_HB - single precision height-balanced, "
+       "DOUBLE_PREC_HB - double precision height-balanced.",
+       SESSION_VAR(histogram_type), CMD_LINE(REQUIRED_ARG),
+       histogram_types, DEFAULT(0));
+
 static Sys_var_mybool Sys_no_thread_alarm(
        "debug_no_thread_alarm",
        "Disable system thread alarm calls. Disabling it may be useful "

sql/table.cc

@@ -425,6 +425,8 @@ void TABLE_SHARE::destroy()
   free_root(&stats_cb.mem_root, MYF(0));
   stats_cb.stats_can_be_read= FALSE;
   stats_cb.stats_is_read= FALSE;
+  stats_cb.histograms_can_be_read= FALSE;
+  stats_cb.histograms_are_read= FALSE;
   if (tmp_table == NO_TMP_TABLE)
     mysql_mutex_unlock(&LOCK_ha_data);
 
@@ -2749,7 +2751,7 @@ partititon_err:
   /* Allocate bitmaps */
 
   bitmap_size= share->column_bitmap_size;
-  if (!(bitmaps= (uchar*) alloc_root(&outparam->mem_root, bitmap_size*5)))
+  if (!(bitmaps= (uchar*) alloc_root(&outparam->mem_root, bitmap_size*6)))
     goto err;
   bitmap_init(&outparam->def_read_set,
               (my_bitmap_map*) bitmaps, share->fields, FALSE);
@@ -2761,8 +2763,12 @@ partititon_err:
               (my_bitmap_map*) (bitmaps+bitmap_size*3), share->fields, FALSE);
   bitmap_init(&outparam->eq_join_set,
               (my_bitmap_map*) (bitmaps+bitmap_size*4), share->fields, FALSE);
+  bitmap_init(&outparam->cond_set,
+              (my_bitmap_map*) (bitmaps+bitmap_size*5), share->fields, FALSE);
   outparam->default_column_bitmaps();
 
+  outparam->cond_selectivity= 1.0;
+
   /* The table struct is now initialized;  Open the table */
   if (db_stat)
   {
@@ -3884,6 +3890,7 @@ void TABLE::init(THD *thd, TABLE_LIST *tl)
   file->ha_start_of_new_statement();
   reginfo.impossible_range= 0;
   created= TRUE;
+  cond_selectivity= 1.0;
 
   /* Catch wrong handling of the auto_increment_field_not_null. */
   DBUG_ASSERT(!auto_increment_field_not_null);
@@ -3892,6 +3899,11 @@ void TABLE::init(THD *thd, TABLE_LIST *tl)
   pos_in_table_list= tl;
 
   clear_column_bitmaps();
+  for (Field **f_ptr= field ; *f_ptr ; f_ptr++)
+  {
+    (*f_ptr)->next_equal_field= NULL;
+    (*f_ptr)->cond_selectivity= 1.0;
+  }
 
   DBUG_ASSERT(key_read == 0);
 

sql/table.h

@@ -586,7 +586,9 @@ struct TABLE_STATISTICS_CB
   Table_statistics *table_stats; /* Structure to access the statistical data */
   bool stats_can_be_read;        /* Memory for statistical data is allocated */
   bool stats_is_read;            /* Statistical data for table has been read
-                                    from statistical tables */   
+                                    from statistical tables */
+  bool histograms_can_be_read;
+  bool histograms_are_read;   
 };
 
 
@@ -1107,6 +1109,7 @@ public:
   my_bitmap_map	*bitmap_init_value;
   MY_BITMAP     def_read_set, def_write_set, def_vcol_set, tmp_set; 
   MY_BITMAP     eq_join_set;         /* used to mark equi-joined fields */
+  MY_BITMAP     cond_set;   /* used to mark fields from sargable conditions*/
   MY_BITMAP     *read_set, *write_set, *vcol_set; /* Active column sets */
   /*
    The ID of the query that opened and is using this table. Has different
@@ -1159,6 +1162,8 @@ public:
   */
   ha_rows       quick_condition_rows;
 
+  double cond_selectivity;
+
   table_map	map;                    /* ID bit of table (1,2,4,8,16...) */
 
   uint          lock_position;          /* Position in MYSQL_LOCK.table */
@@ -1278,6 +1283,7 @@ public:
 #endif
   uint max_keys; /* Size of allocated key_info array. */
   bool stats_is_read;     /* Persistent statistics is read for the table */
+  bool histograms_are_read;
   MDL_ticket *mdl_ticket;
 
   void init(THD *thd, TABLE_LIST *tl);

sql/uniques.cc

@@ -86,6 +86,7 @@ Unique::Unique(qsort_cmp2 comp_func, void * comp_func_fixed_arg,
   full_size= size;
   if (min_dupl_count_arg)
     full_size+= sizeof(element_count);
+  with_counters= test(min_dupl_count_arg);
   my_b_clear(&file);
   init_tree(&tree, (ulong) (max_in_memory_size / 16), 0, size, comp_func,
             NULL, comp_func_fixed_arg, MYF(MY_THREAD_SPECIFIC));
@@ -428,6 +429,22 @@ static int buffpek_compare(void *arg, uchar *key_ptr1, uchar *key_ptr2)
 C_MODE_END
 
 
+inline
+element_count get_counter_from_merged_element(void *ptr, uint ofs)
+{
+  element_count cnt;
+  memcpy((uchar *) &cnt, (uchar *) ptr + ofs, sizeof(element_count));
+  return cnt;
+}
+
+
+inline
+void put_counter_into_merged_element(void *ptr, uint ofs, element_count cnt)
+{
+  memcpy((uchar *) ptr + ofs, (uchar *) &cnt, sizeof(element_count));
+}
+
+
 /*
   DESCRIPTION
 
@@ -457,6 +474,8 @@ C_MODE_END
     file               file with all trees dumped. Trees in the file
                        must contain sorted unique values. Cache must be
                        initialized in read mode.
+    with counters      take into account counters for equal merged
+                       elements
   RETURN VALUE
     0     ok
     <> 0  error
@@ -466,7 +485,7 @@ static bool merge_walk(uchar *merge_buffer, ulong merge_buffer_size,
                        uint key_length, BUFFPEK *begin, BUFFPEK *end,
                        tree_walk_action walk_action, void *walk_action_arg,
                        qsort_cmp2 compare, void *compare_arg,
-                       IO_CACHE *file)
+                       IO_CACHE *file, bool with_counters)
 {
   BUFFPEK_COMPARE_CONTEXT compare_context = { compare, compare_arg };
   QUEUE queue;
@@ -485,6 +504,8 @@ static bool merge_walk(uchar *merge_buffer, ulong merge_buffer_size,
   uint bytes_read;               /* to hold return value of read_to_buffer */
   BUFFPEK *top;
   int res= 1;
+  uint cnt_ofs= key_length - (with_counters ? sizeof(element_count) : 0);
+  element_count cnt;
   /*
     Invariant: queue must contain top element from each tree, until a tree
     is not completely walked through.
@@ -543,9 +564,17 @@ static bool merge_walk(uchar *merge_buffer, ulong merge_buffer_size,
     /* new top has been obtained; if old top is unique, apply the action */
     if (compare(compare_arg, old_key, top->key))
     {
-      if (walk_action(old_key, 1, walk_action_arg))
+      cnt= with_counters ?
+           get_counter_from_merged_element(old_key, cnt_ofs) : 1;
+      if (walk_action(old_key, cnt, walk_action_arg))
         goto end;
     }
+    else if (with_counters)
+    {
+      cnt= get_counter_from_merged_element(top->key, cnt_ofs);
+      cnt+= get_counter_from_merged_element(old_key, cnt_ofs);
+      put_counter_into_merged_element(top->key, cnt_ofs, cnt);
+    }
   }
   /*
     Applying walk_action to the tail of the last tree: this is safe because
@@ -556,7 +585,10 @@ static bool merge_walk(uchar *merge_buffer, ulong merge_buffer_size,
   {
     do
     {
-      if (walk_action(top->key, 1, walk_action_arg))
+      
+      cnt= with_counters ?
+           get_counter_from_merged_element(top->key, cnt_ofs) : 1;
+      if (walk_action(top->key, cnt, walk_action_arg))
         goto end;
       top->key+= key_length;
     }
@@ -620,7 +652,7 @@ bool Unique::walk(TABLE *table, tree_walk_action action, void *walk_action_arg)
                     (BUFFPEK *) file_ptrs.buffer,
                     (BUFFPEK *) file_ptrs.buffer + file_ptrs.elements,
                     action, walk_action_arg,
-                    tree.compare, tree.custom_arg, &file);
+                    tree.compare, tree.custom_arg, &file, with_counters);
   }
   my_free(merge_buffer);
   return res;