mariadb/mysql-test/main/analyze_format_json.test
Sergei Petrunia e87d1e391b MDEV-18478 ANALYZE for statement should show selectivity of ICP, part#1
(Based on the original patch by Jason Cu)

Part #1:
- Add ha_handler_stats::{icp_attempts,icp_match}, make
  handler_index_cond_check() increment them.
- ANALYZE FORMAT=JSON now prints r_icp_filtered based on these counters.
2024-04-23 22:55:22 +03:00

348 lines
11 KiB
Text

#
# Tests for "ANALYZE FORMAT=JSON $statement" syntax
#
--disable_warnings
drop table if exists t0,t1,t2,t3;
--enable_warnings
--source include/default_optimizer_switch.inc
create table t0 (a int);
INSERT INTO t0 VALUES (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
--echo # r_filtered=30%, because 3 rows match: 0,1,2
--source include/analyze-format.inc
analyze format=json select * from t0 where a<3;
create table t1 (a int, b int, c int, key(a));
insert into t1 select A.a*10 + B.a, A.a*10 + B.a, A.a*10 + B.a from t0 A, t0 B;
--source include/analyze-format.inc
analyze
select * from t0, t1 where t1.a=t0.a and t0.a > 9;
--source include/analyze-format.inc
analyze format=json
select * from t0, t1 where t1.a=t0.a and t0.a > 9;
analyze
select * from t0, t1 where t1.a=t0.a and t1.b<4;
--source include/analyze-format.inc
analyze format=json
select * from t0, t1 where t1.a=t0.a and t1.b<4;
analyze
select * from t1 tbl1, t1 tbl2 where tbl1.b<2 and tbl2.b>5;
--source include/analyze-format.inc
analyze format=json
select * from t1 tbl1, t1 tbl2 where tbl1.b<20 and tbl2.b<60;
--source include/analyze-format.inc
analyze format=json
select * from t1 tbl1, t1 tbl2 where tbl1.b<20 and tbl2.b<60 and tbl1.c > tbl2.c;
drop table t1;
drop table t0;
--echo #
--echo # MDEV-7674: ANALYZE shows r_rows=0
--echo #
create table t1(a int);
insert into t1 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
create table t2 (a int, key(a));
insert into t2 values (0),(1);
--source include/analyze-format.inc
analyze format=json select * from t1 straight_join t2 force index(a) where t2.a=t1.a;
drop table t1,t2;
create table t1(a int);
insert into t1 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
--echo #
--echo # MDEV-7679: ANALYZE crashes when printing WHERE when no default db
--echo #
select database();
connect (con1,localhost,root,,*NO-ONE*);
connection con1;
select database();
--source include/analyze-format.inc
analyze format=json select * from test.t1 where t1.a<5;
disconnect con1;
connection default;
drop table t1;
--echo #
--echo # MDEV-7812: ANALYZE FORMAT=JSON UPDATE/DELETE doesnt print
--echo # the r_total_time_ms
--echo #
create table t2(a int);
insert into t2 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
create table t3(a int);
insert into t3 select A.a + B.a* 10 + C.a * 100 from t2 A, t2 B, t2 C;
create table t1 (pk int primary key);
insert into t1 select a from t3;
alter table t1 add b int;
--source include/analyze-format.inc
analyze format=json
update t1 set b=pk;
--source include/analyze-format.inc
analyze format=json
select * from t1 where pk < 10 and b > 4;
--source include/analyze-format.inc
analyze format=json
delete from t1 where pk < 10 and b > 4;
drop table t1, t3, t2;
--echo #
--echo # MDEV-7833:ANALYZE FORMAT=JSON and Range checked for each record
--echo #
create table t3(a int);
insert into t3 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
create table t4(a int);
insert into t4 select A.a + B.a* 10 + C.a * 100 from t3 A, t3 B, t3 C;
create table t1 (lb1 int, rb1 int, lb2 int, rb2 int, c1 int, c2 int);
insert into t1 values (1,2,10,20,15,15);
insert into t1 values (3,5,10,20,15,15);
insert into t1 values (10,20,10,20,15,15);
insert into t1 values (10,20,1,2,15,15);
insert into t1 values (10,20,10,20,1,3);
create table t2 (key1 int, key2 int, key3 int, key4 int, col1 int,
key(key1), key(key2), key(key3), key(key4));
insert into t2 select a,a,a,a,a from t3;
insert into t2 select 15,15,15,15,15 from t4;
--source include/analyze-format.inc
analyze format=json
select * from t1, t2 where (t2.key1 between t1.lb1 and t1.rb1) and
(t2.key2 between t1.lb2 and t1.rb2) and
(t2.key3=t1.c1 OR t2.key4=t1.c2);
drop table t1,t2,t3,t4;
--echo #
--echo # MDEV-7904: ANALYZE FORMAT=JSON SELECT .. UNION SELECT doesn't print r_rows for union output
--echo #
create table t0 (a int);
INSERT INTO t0 VALUES (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
create table t1 (a int);
INSERT INTO t1 select * from t0;
--source include/analyze-format.inc
analyze format=json (select * from t1 tbl1 where a<5) union (select * from t1 tbl2 where a in (2,3));
drop table t0, t1;
--echo #
--echo # MDEV-7970: EXPLAIN FORMAT=JSON does not print HAVING
--source include/explain-no-costs.inc
--echo #
create table t0(a int);
insert into t0 values (0),(1),(2),(3);
create table t1(a int);
insert into t1 select A.a + B.a* 10 + C.a * 100 from t0 A, t0 B, t0 C;
create table t2 (
a int,
b int,
key (a)
);
insert into t2 select A.a*1000 + B.a, A.a*1000 + B.a from t0 A, t1 B;
--echo # normal HAVING
--source include/analyze-format.inc
analyze format=json select a, max(b) as TOP from t2 group by a having TOP > a;
--echo # HAVING is always TRUE (not printed)
--source include/analyze-format.inc
analyze format=json select a, max(b) as TOP from t2 group by a having 1<>2;
--echo # HAVING is always FALSE (intercepted by message)
--source include/analyze-format.inc
analyze format=json select a, max(b) as TOP from t2 group by a having 1=2;
--echo # HAVING is absent
--source include/analyze-format.inc
analyze format=json select a, max(b) as TOP from t2 group by a;
drop table t0, t1, t2;
--echo #
--echo # MDEV-7267: Server crashes in Item_field::print on ANALYZE FORMAT=JSON
--echo #
CREATE TABLE t1 (a INT);
INSERT INTO t1 VALUES (1),(2);
CREATE TABLE t2 (b INT);
INSERT INTO t2 VALUES (3),(4);
--source include/analyze-format.inc
ANALYZE FORMAT=JSON SELECT STRAIGHT_JOIN * FROM t1, t2 WHERE b IN ( SELECT a FROM t1 );
SELECT STRAIGHT_JOIN * FROM t1, t2 WHERE b IN ( SELECT a FROM t1 );
drop table t1,t2;
--echo #
--echo # MDEV-8864: Server crash #2 in Item_field::print on ANALYZE FORMAT=JSON
--echo #
CREATE TABLE t1 (f1 INT) ENGINE=MyISAM;
INSERT INTO t1 VALUES (1),(2);
CREATE TABLE t2 (f2 INT) ENGINE=MyISAM;
INSERT INTO t2 VALUES (2),(3);
CREATE TABLE t3 (f3 INT) ENGINE=MyISAM;
INSERT INTO t3 VALUES (3),(4);
--source include/analyze-format.inc
ANALYZE FORMAT=JSON
SELECT GROUP_CONCAT(f3) AS gc, ( SELECT MAX(f1) FROM t1, t2 WHERE f2 = f3 ) sq
FROM t2, t3
WHERE f3 IN ( 1, 2 )
GROUP BY sq ORDER BY gc;
drop table t1,t2,t3;
--echo #
--echo # MDEV-13286: Floating point exception in Filesort_tracker::print_json_members(Json_writer*)
--echo #
create table t0(a int);
insert into t0 values (0),(1),(2),(3),(4),(5),(6),(7),(8),(9);
create table t1 (a int, b int, c int);
insert into t1 select a,a,a from t0;
create table t2 as select * from t1;
--source include/analyze-format.inc
analyze format=json select a, (select t2.b from t2 where t2.a<t1.a order by t2.c limit 1) from t1 where t1.a<0;
drop table t0,t1,t2;
--echo #
--echo # MDEV-30806: ANALYZE FORMAT=JSON: better support for BNL and BNL-H joins
--echo #
--source include/have_sequence.inc
create table t10 (
a int,
b int
);
insert into t10 select seq, seq/3 from seq_0_to_999;
create table t11 (
a int,
b int
);
insert into t11 select seq, seq/5 from seq_0_to_999;
analyze table t10,t11 persistent for all;
--source include/analyze-format.inc
analyze format=json
select * from t10, t11
where
t10.a < 700 and
t11.a < 100
and t10.b=t11.b;
set @tmp=@@join_cache_level, join_cache_level=6;
--source include/analyze-format.inc
analyze format=json
select * from t10, t11
where
t10.a < 700 and
t11.a < 100
and t10.b=t11.b;
set join_cache_level=@tmp;
drop table t10, t11;
--echo #
--echo # MDEV-18478: ANALYZE for statement should show selectivity of
--echo # pushed index condition
--echo #
create table t12 (a int, b varchar(10), c int, index(a,b), index(c));
create table t13 (a int not null primary key, b int, index(b));
insert into t12(a,b,c) values (1,"abc",100),
(2,"abd",200),(2,"bfd",300),(2,"efg",400),
(3,"abc",410),(3,"def",420),(3,"ghi",430),(3,"jkl",440),
(4,"abc",600),(4,"def",700),(4,"ghi",800),(4,"jkl",900);
insert into t13(a,b) values(1,1),(2,2),(3,3),(4,4),(5,5);
analyze table t12;
analyze table t13;
-- echo # eq key condition only, no pushed condition
# no r_pushed_condition in output because there is no pushed condition
# a=2 qualifies 3 rows (r_rows=3)
--source include/analyze-format.inc
analyze format=json select * from t12 where t12.a=2;
-- echo # pushed index condition
# key condition a=2 gets 3 rows
# pushed condition b like '%f%' filters down to two of the 3 rows (66.667%)
--source include/analyze-format.inc
analyze format=json select * from t12 where t12.a=2 and t12.b like '%f%';
-- echo # rowid filter only, no pushed condition
# key condition a=2 gets 3 rows
# rowid filter condition between 400 and 500 qualifies 1 of the 3 key rows (33.33%)
--source include/analyze-format.inc
analyze format=json select * from t12 where t12.a=2 and t12.c
between 400 and 500;
-- echo # pushed index condition and rowid filter
# key condition a=2 gets 3 rows
# pushed condition b like '%f%' filters down to 2 of the 3 key rows (66.667%)
# rowid filter condition between 400 and 500 filters down to 1 of the icp 2 rows (50%)
--source include/analyze-format.inc
analyze format=json select * from t12 where t12.a=2 and t12.c
between 400 and 500 and t12.b like '%f%';
-- echo # pushed index condition on the inner table (t12) of a join
# for inner table t12, key t12.a=t13.a
# for t13.a=1, 1 matching row in t12, 100% filtered by the index condition (t12.b like '%f%')
# for t13.a=2, 3 matching rows in t12, 66% filtered by the index cond
# r_rows_idx=2 (average 4 (1+3)/2 scans = 2 rows per scan)
# r_icp_filtered=50 (4 rows total, 2 survive the icp, both in the second scan)
--source include/analyze-format.inc
analyze format=json select * from t13,t12
where t13.a between 1 and 2 and t12.a=t13.a and t12.b like '%f%';
-- echo # rowid filter on inner table of a join, no ICP
# inner table t12, key t12.a=t13.a
# for t13.a=1, 1 matching row in t12, 0% survive the rowid filter (t12.c between 400 and 500)
# for t13.a=2, 3 matching rows in t12, 33% (1/3) survive the rowid filter
# rowid_filter r_lookups=2 (1+3 / 2 loops)
# rowid_filter r_selectivity_pct = 25 (1 / (1+3))
--source include/analyze-format.inc
analyze format=json select * from t13,t12 where t13.b between 1 and 2 and t12.a=t13.a and t12.c
between 400 and 450;
-- echo # rowid filter and ICP on inner table of a join
# inner table t12, key t12.a=t13.a
# for t13.a=1, 1 matching row in t12, 100% filtered by the index condition (t12.b like '%f%')
# for t13.a=2, 3 matching rows in t12, 66% filtered by the index cond
# r_rows_idx=2 (average 4 (1+3)/2 scans = 2 rows per scan)
# r_icp_filtered=50 (4 rows total, 2 survive the icp, both in the second scan)
# for t13.a=1, 0 rows after ICP
# for t13.a=2, 2 rows after ICP, 1 survives the rowid_filter (t12.c bewteen 400 and 500)
# rowid_filter r_lookups=2 (2 / 2 loops)
# rowid_filter r_selectivity_pct = 50 (1 survived / 2 rows after ICP)
--source include/analyze-format.inc
analyze format=json select * from t13,t12 where t13.b between 1 and 2 and t12.a=t13.a and t12.c
between 400 and 450 and t12.b like '%f%';
drop table t12,t13;