Skip to content

M
mariadb
Commit bd305d6b authored by Igor Babaev's avatar Igor Babaev
Browse files
Options

Fixed bug mdev-4220. 

This bug is a regression bug. The regression was introduced by
the patch for mdev-3851, that tried to weaken the condition when
a ref access with an extended key can be converted to an eq_ref
access. The patch incorrectly formed this condition. As a result,
while improving performance for some queries, the patch caused 
worse performance for another queries.
parent 9d72bbf8
Hide whitespace changes
Inline Side-by-side
Showing with 181 additions and 16 deletions
mysql-test/r/innodb_ext_key.result
View file @ bd305d6b
......@@ -731,6 +731,7 @@ CREATE TABLE t2 (b int) ENGINE=MyISAM;
INSERT INTO t1 (a) VALUES (4), (6);
INSERT INTO t2 (b) VALUES (0), (8);
set @save_optimizer_switch=@@optimizer_switch;
set @save_join_cache_level=@@join_cache_level;
SET join_cache_level=3;
SET optimizer_switch='join_cache_hashed=on';
SET optimizer_switch='join_cache_bka=on';
......@@ -742,6 +743,7 @@ id select_type table type possible_keys key key_len ref rows Extra
1 SIMPLE t2 hash_ALL NULL #hash#$hj 5 test.t1.a 2 Using where; Using join buffer (flat, BNLH join)
SELECT * FROM t1, t2 WHERE b=a;
a b
set join_cache_level=@save_join_cache_level;
set optimizer_switch=@save_optimizer_switch;
DROP TABLE t1,t2;
#
......@@ -772,5 +774,97 @@ INSERT INTO t1 (c2, c3, c4) VALUES (58291525, 2580, 'foobar')
ON DUPLICATE KEY UPDATE c4 = VALUES(c4);
set optimizer_switch=@save_optimizer_switch;
DROP TABLE t1;
#
# Bug mdev-4220: using ref instead of eq_ref
# with extended_keys=on
# (performance regression introduced in the patch for mdev-3851)
#
set @save_optimizer_switch=@@optimizer_switch;
create table t1 (a int not null) engine=innodb;
insert into t1 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
create table t2 (
pk int primary key, a int not null, b int, unique(a)
)engine=innodb;
insert into t2
select
A.a + 10 * B.a, A.a + 10 * B.a, A.a + 10 * B.a
from t1 A, t1 B;
set optimizer_switch='extended_keys=off';
explain
select * from t1, t2 where t2.a=t1.a and t2.b < 2;
id select_type table type possible_keys key key_len ref rows Extra
1 SIMPLE t1 ALL NULL NULL NULL NULL 10
1 SIMPLE t2 eq_ref a a 4 test.t1.a 1 Using where
flush status;
select * from t1, t2 where t2.a=t1.a and t2.b < 2;
a pk a b
0 0 0 0
1 1 1 1
show status like 'handler_read%';
Variable_name Value
Handler_read_first 0
Handler_read_key 10
Handler_read_last 0
Handler_read_next 0
Handler_read_prev 0
Handler_read_rnd 0
Handler_read_rnd_deleted 0
Handler_read_rnd_next 11
set optimizer_switch='extended_keys=on';
explain
select * from t1, t2 where t2.a=t1.a and t2.b < 2;
id select_type table type possible_keys key key_len ref rows Extra
1 SIMPLE t1 ALL NULL NULL NULL NULL 10
1 SIMPLE t2 eq_ref a a 4 test.t1.a 1 Using where
flush status;
select * from t1, t2 where t2.a=t1.a and t2.b < 2;
a pk a b
0 0 0 0
1 1 1 1
show status like 'handler_read%';
Variable_name Value
Handler_read_first 0
Handler_read_key 10
Handler_read_last 0
Handler_read_next 0
Handler_read_prev 0
Handler_read_rnd 0
Handler_read_rnd_deleted 0
Handler_read_rnd_next 11
drop table t1,t2;
create table t1(a int) engine=myisam;
insert into t1 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
create table t2(a int) engine=myisam;
insert into t2 select A.a + 10*B.a + 100*C.a from t1 A, t1 B, t1 C;
create table t3 (
pk1 int not null, pk2 int not null, col1 int not null, col2 int not null)
engine=innodb;
insert into t3 select a,a,a,a from t2;
alter table t3 add primary key (pk1, pk2);
alter table t3 add key (col1, col2);
set optimizer_switch='extended_keys=off';
explain
select * from t1, t3 where t3.col1=t1.a and t3.col2=t1.a;
id select_type table type possible_keys key key_len ref rows Extra
1 SIMPLE t1 ALL NULL NULL NULL NULL # Using where
1 SIMPLE t3 ref col1 col1 8 test.t1.a,test.t1.a # Using index
explain
select * from t1, t3 where t3.col1=t1.a and t3.col2=t1.a and t3.pk1=t1.a;
id select_type table type possible_keys key key_len ref rows Extra
1 SIMPLE t1 ALL NULL NULL NULL NULL # Using where
1 SIMPLE t3 ref PRIMARY,col1 PRIMARY 4 test.t1.a # Using where
set optimizer_switch='extended_keys=on';
explain
select * from t1, t3 where t3.col1=t1.a and t3.col2=t1.a;
id select_type table type possible_keys key key_len ref rows Extra
1 SIMPLE t1 ALL NULL NULL NULL NULL # Using where
1 SIMPLE t3 ref col1 col1 8 test.t1.a,test.t1.a # Using index
explain
select * from t1, t3 where t3.col1=t1.a and t3.col2=t1.a and t3.pk1=t1.a;
id select_type table type possible_keys key key_len ref rows Extra
1 SIMPLE t1 ALL NULL NULL NULL NULL # Using where
1 SIMPLE t3 ref PRIMARY,col1 col1 12 test.t1.a,test.t1.a,test.t1.a # Using index
drop table t1,t2,t3;
set optimizer_switch=@save_optimizer_switch;
set optimizer_switch=@save_ext_key_optimizer_switch;
SET SESSION STORAGE_ENGINE=DEFAULT;
mysql-test/t/innodb_ext_key.test
View file @ bd305d6b
......@@ -414,6 +414,7 @@ INSERT INTO t1 (a) VALUES (4), (6);
INSERT INTO t2 (b) VALUES (0), (8);
set @save_optimizer_switch=@@optimizer_switch;
set @save_join_cache_level=@@join_cache_level;
SET join_cache_level=3;
SET optimizer_switch='join_cache_hashed=on';
......@@ -424,6 +425,7 @@ EXPLAIN
SELECT * FROM t1, t2 WHERE b=a;
SELECT * FROM t1, t2 WHERE b=a;
set join_cache_level=@save_join_cache_level;
set optimizer_switch=@save_optimizer_switch;
DROP TABLE t1,t2;
......@@ -466,6 +468,78 @@ set optimizer_switch=@save_optimizer_switch;
DROP TABLE t1;
--echo #
--echo # Bug mdev-4220: using ref instead of eq_ref
--echo # with extended_keys=on
--echo # (performance regression introduced in the patch for mdev-3851)
--echo #
set @save_optimizer_switch=@@optimizer_switch;
create table t1 (a int not null) engine=innodb;
insert into t1 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
create table t2 (
pk int primary key, a int not null, b int, unique(a)
)engine=innodb;
insert into t2
select
A.a + 10 * B.a, A.a + 10 * B.a, A.a + 10 * B.a
from t1 A, t1 B;
set optimizer_switch='extended_keys=off';
explain
select * from t1, t2 where t2.a=t1.a and t2.b < 2;
flush status;
select * from t1, t2 where t2.a=t1.a and t2.b < 2;
show status like 'handler_read%';
set optimizer_switch='extended_keys=on';
explain
select * from t1, t2 where t2.a=t1.a and t2.b < 2;
flush status;
select * from t1, t2 where t2.a=t1.a and t2.b < 2;
show status like 'handler_read%';
drop table t1,t2;
# this test case did not demonstrate any regression
# it is added for better testing
create table t1(a int) engine=myisam;
insert into t1 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
create table t2(a int) engine=myisam;
insert into t2 select A.a + 10*B.a + 100*C.a from t1 A, t1 B, t1 C;
create table t3 (
pk1 int not null, pk2 int not null, col1 int not null, col2 int not null)
engine=innodb;
insert into t3 select a,a,a,a from t2;
alter table t3 add primary key (pk1, pk2);
alter table t3 add key (col1, col2);
set optimizer_switch='extended_keys=off';
--replace_column 9 #
explain
select * from t1, t3 where t3.col1=t1.a and t3.col2=t1.a;
--replace_column 9 #
explain
select * from t1, t3 where t3.col1=t1.a and t3.col2=t1.a and t3.pk1=t1.a;
set optimizer_switch='extended_keys=on';
--replace_column 9 #
explain
select * from t1, t3 where t3.col1=t1.a and t3.col2=t1.a;
--replace_column 9 #
explain
select * from t1, t3 where t3.col1=t1.a and t3.col2=t1.a and t3.pk1=t1.a;
drop table t1,t2,t3;
set optimizer_switch=@save_optimizer_switch;
set optimizer_switch=@save_ext_key_optimizer_switch;
SET SESSION STORAGE_ENGINE=DEFAULT;
sql/sql_select.cc
View file @ bd305d6b
......@@ -7987,20 +7987,17 @@ static bool create_ref_for_key(JOIN *join, JOIN_TAB *j,
ulong key_flags= j->table->actual_key_flags(keyinfo);
if (j->type == JT_CONST)
j->table->const_table= 1;
else if (((key_flags & (HA_NOSAME | HA_NULL_PART_KEY))!= HA_NOSAME) ||
keyparts != j->table->actual_n_key_parts(keyinfo) ||
null_ref_key)
{
if (test(key_flags & HA_EXT_NOSAME) && keyparts == keyinfo->ext_key_parts &&
!null_ref_key)
j->type= JT_EQ_REF;
else
{
/* Must read with repeat */
j->type= null_ref_key ? JT_REF_OR_NULL : JT_REF;
j->ref.null_ref_key= null_ref_key;
j->ref.null_ref_part= null_ref_part;
}
else if (!((keyparts == keyinfo->key_parts &&
((key_flags & (HA_NOSAME | HA_NULL_PART_KEY)) == HA_NOSAME)) ||
(keyparts > keyinfo->key_parts && // true only for extended keys
test(key_flags & HA_EXT_NOSAME) &&
keyparts == keyinfo->ext_key_parts)) ||
null_ref_key)
{
/* Must read with repeat */
j->type= null_ref_key ? JT_REF_OR_NULL : JT_REF;
j->ref.null_ref_key= null_ref_key;
j->ref.null_ref_part= null_ref_part;
}
else if (keyuse_uses_no_tables)
{
......
sql/table.cc
View file @ bd305d6b
......@@ -1003,7 +1003,7 @@ static int open_binary_frm(THD *thd, TABLE_SHARE *share, uchar *head,
}
}
if (j == first_key_parts)
keyinfo->ext_key_flags= keyinfo->flags | HA_NOSAME | HA_EXT_NOSAME;
keyinfo->ext_key_flags= keyinfo->flags | HA_EXT_NOSAME;
}
share->ext_key_parts+= keyinfo->ext_key_parts;
}
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7