Skip to content

MDEV-37936 Followup to MDEV-36321: out_rows for GROUP BY: use of item… #4415

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
+232 −47
Open

MDEV-37936 Followup to MDEV-36321: out_rows for GROUP BY: use of item… #4415

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
1ff90c2
MDEV-37936 Followup to MDEV-36321: out_rows for GROUP BY: use of item…
mariadb-RexJohnston Oct 31, 2025
274ccd3
Merge branch 'main' into 12.2-MDEV-37936-v2
mariadb-RexJohnston Nov 11, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
101 changes: 92 additions & 9 deletions mysql-test/main/derived_opt.result
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -595,7 +595,7 @@ select * from
t2,
(select max(value), grp_id from t1 group by grp_id) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;
a max(value) grp_id
1 100 1
2 100 2
Expand All @@ -616,14 +616,38 @@ t
"rec_per_key_estimate": 1
}
]
explain
with DT as
(
select max(value), 1/grp_id as fn from t1 group by 2
)
select * from t2, DT where t2.a = DT.fn;
id select_type table type possible_keys key key_len ref rows Extra
1 PRIMARY t2 ALL NULL NULL NULL NULL 5 Using where
1 PRIMARY <derived2> ref key0 key0 4 test.t2.a 1 Using where
2 DERIVED t1 ALL NULL NULL NULL NULL 10000 Using temporary; Using filesort
select
json_detailed(json_extract(trace, '$**.infer_derived_key_statistics')) as t
from information_schema.optimizer_trace;
t
[
{
"table_alias": "DT",
"key_name": "key0",
"key_parts": 1,
"select":
["group_list_in_key"],
"rec_per_key_estimate": 1
}
]
# Same as above, but try a UNION:
select * from
t2,
(select max(value), grp_id from t1 group by grp_id
union all
select max(value), grp_id from t1 group by grp_id) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;
a max(value) grp_id
1 100 1
1 100 1
Expand Down Expand Up @@ -660,7 +684,7 @@ t2,
union all
select max(value), grp_id from t1 group by MOD(grp_id,2)) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;
id select_type table type possible_keys key key_len ref rows Extra
1 PRIMARY t2 ALL NULL NULL NULL NULL 5 Using where
1 PRIMARY <derived2> ref key0 key0 5 test.t2.a 101
Expand Down Expand Up @@ -689,7 +713,7 @@ select * from
t2,
(select grp_id, max(value) as maxval from v1 group by grp_id) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;
id select_type table type possible_keys key key_len ref rows Extra
1 PRIMARY t2 ALL NULL NULL NULL NULL 5 Using where
1 PRIMARY <derived2> ref key0 key0 5 test.t2.a 1
Expand All @@ -701,7 +725,7 @@ select * from
t2,
(select grp_id, max(value) as maxval from cte1 group by grp_id) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;
id select_type table type possible_keys key key_len ref rows Extra
1 PRIMARY t2 ALL NULL NULL NULL NULL 5 Using where
1 PRIMARY <derived3> ref key0 key0 5 test.t2.a 1
Expand All @@ -718,7 +742,7 @@ where t1.grp_id = t3.b
group by grp_id
) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;
id select_type table type possible_keys key key_len ref rows Extra
1 PRIMARY t2 ALL NULL NULL NULL NULL 5 Using where
1 PRIMARY <derived2> ref key0 key0 5 test.t2.a 1
Expand All @@ -733,7 +757,7 @@ t2,
(select max(value) as maxval, grp_id from t1 group by grp_id) DT
where
t2.col2=maxval and
t2.a= DT.grp_id;
t2.a = DT.grp_id;
id select_type table type possible_keys key key_len ref rows Extra
1 PRIMARY t2 ALL NULL NULL NULL NULL 5 Using where
1 PRIMARY <derived2> ref key0 key0 10 test.t2.col2,test.t2.a 1
Expand All @@ -757,12 +781,29 @@ select * from
t2,
(select grp_id, max(value) as maxval from t1 group by grp_id) DT
where
t2.col2=maxval and
t2.a= DT.grp_id;
t2.col2 = maxval and
t2.a = DT.grp_id;
id select_type table type possible_keys key key_len ref rows Extra
1 PRIMARY t2 ALL NULL NULL NULL NULL 5 Using where
1 PRIMARY <derived2> ref key0 key0 10 test.t2.a,test.t2.col2 1
2 DERIVED t1 ALL grp_id NULL NULL NULL 10000 Using temporary; Using filesort
explain
select * from
t2 join
(
select max(value) as X, grp_id from t1 group by grp_id
union all
select distinct 'Total' as X, b as grp_id from t3
) DT
on t2.a = DT.grp_id
where DT.X = 100;
id select_type table type possible_keys key key_len ref rows Extra
1 PRIMARY t2 ALL NULL NULL NULL NULL 5 Using where
1 PRIMARY <derived2> ref key0 key0 5 test.t2.a 2 Using where
2 DERIVED t1 ALL NULL NULL NULL NULL 10000 Using temporary; Using filesort
3 UNION NULL NULL NULL NULL NULL NULL NULL Impossible WHERE
Warnings:
Warning 1292 Truncated incorrect DECIMAL value: 'Total'
delete from t1;
insert into t1 select 1, a.seq from seq_1_to_10 a;
analyze table t1;
Expand Down Expand Up @@ -1060,6 +1101,48 @@ t
}
]
drop table t1, t2, t3, t4;
create table t1 (
grp_id int,
value int,
index (grp_id)
);
insert into t1 select
A.seq, B.seq
from
seq_1_to_100 A,
seq_1_to_100 B;
create table t1a(
grp_id_2 int,
value int,
index (grp_id_2)
);
insert into t1a select * from t1;
create table t2 (a int);
insert into t2 select seq from seq_1_to_5;
analyze table t1,t1a, t2;
Table Op Msg_type Msg_text
test.t1 analyze status Engine-independent statistics collected
test.t1 analyze status Table is already up to date
test.t1a analyze status Engine-independent statistics collected
test.t1a analyze status Table is already up to date
test.t2 analyze status Engine-independent statistics collected
test.t2 analyze status OK
# <derived2> must have type=ref, rows=2 (not 10):
explain
select * from
t2,
(
select max(value), grp_id from t1 group by grp_id
union all
select max(value), grp_id_2 from t1a group by grp_id_2
) DT
where t2.a= DT.grp_id;
id select_type table type possible_keys key key_len ref rows Extra
1 PRIMARY t2 ALL NULL NULL NULL NULL 5 Using where
1 PRIMARY <derived2> ref key0 key0 5 test.t2.a 2
2 DERIVED t1 ALL NULL NULL NULL NULL 10000 Using temporary; Using filesort
3 UNION t1a ALL NULL NULL NULL NULL 10000 Using temporary; Using filesort
drop table t1, t1a, t2;
#
# End of 11.4 tests
#
Expand Down
81 changes: 72 additions & 9 deletions mysql-test/main/derived_opt.test
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -473,12 +473,22 @@ select * from
t2,
(select max(value), grp_id from t1 group by grp_id) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;

select
json_detailed(json_extract(trace, '$**.infer_derived_key_statistics')) as t
from information_schema.optimizer_trace;

explain
with DT as
(
select max(value), 1/grp_id as fn from t1 group by 2
)
select * from t2, DT where t2.a = DT.fn;

select
json_detailed(json_extract(trace, '$**.infer_derived_key_statistics')) as t
from information_schema.optimizer_trace;

--echo # Same as above, but try a UNION:
select * from
Expand All @@ -487,7 +497,7 @@ select * from
union all
select max(value), grp_id from t1 group by grp_id) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;
select
json_detailed(json_extract(trace, '$**.infer_derived_key_statistics')) as t
from information_schema.optimizer_trace;
Expand All @@ -500,7 +510,7 @@ select * from
union all
select max(value), grp_id from t1 group by MOD(grp_id,2)) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;
select
json_detailed(json_extract(trace, '$**.infer_derived_key_statistics')) as t
from information_schema.optimizer_trace;
Expand All @@ -514,7 +524,7 @@ select * from
t2,
(select grp_id, max(value) as maxval from v1 group by grp_id) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;

drop view v1;

Expand All @@ -524,7 +534,7 @@ select * from
t2,
(select grp_id, max(value) as maxval from cte1 group by grp_id) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;

explain
select * from
Expand All @@ -538,7 +548,7 @@ select * from
group by grp_id
) DT
where
t2.a= DT.grp_id;
t2.a = DT.grp_id;

--echo # Example with equalities on GROUP BY columns and other columns
--echo # Must produce {table=<derived2>, ref=test.t2.col2,test.t2.a, rows=1}
Expand All @@ -549,7 +559,7 @@ select * from
(select max(value) as maxval, grp_id from t1 group by grp_id) DT
where
t2.col2=maxval and
t2.a= DT.grp_id;
t2.a = DT.grp_id;
select
json_detailed(json_extract(trace, '$**.infer_derived_key_statistics')) as t
from information_schema.optimizer_trace;
Expand All @@ -561,9 +571,21 @@ select * from
t2,
(select grp_id, max(value) as maxval from t1 group by grp_id) DT
where
t2.col2=maxval and
t2.a= DT.grp_id;
t2.col2 = maxval and
t2.a = DT.grp_id;

# room for improvement here, 'Total' != 100, will, ignoring the impossible
# condition in the 2nd select, we should see 2 rows per key in DT
explain
select * from
t2 join
(
select max(value) as X, grp_id from t1 group by grp_id
union all
select distinct 'Total' as X, b as grp_id from t3
) DT
on t2.a = DT.grp_id
where DT.X = 100;

delete from t1;
insert into t1 select 1, a.seq from seq_1_to_10 a;
Expand Down Expand Up @@ -709,6 +731,47 @@ from information_schema.optimizer_trace;

drop table t1, t2, t3, t4;

##
## Tests for item names
##

create table t1 (
grp_id int,
value int,
index (grp_id)
);

insert into t1 select
A.seq, B.seq
from
seq_1_to_100 A,
seq_1_to_100 B;

create table t1a(
grp_id_2 int,
value int,
index (grp_id_2)
);
insert into t1a select * from t1;

create table t2 (a int);
insert into t2 select seq from seq_1_to_5;

analyze table t1,t1a, t2;

--echo # <derived2> must have type=ref, rows=2 (not 10):
explain
select * from
t2,
(
select max(value), grp_id from t1 group by grp_id
union all
select max(value), grp_id_2 from t1a group by grp_id_2
) DT
where t2.a= DT.grp_id;

drop table t1, t1a, t2;

--echo #
--echo # End of 11.4 tests
--echo #
Expand Down
Loading