mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-01-27 01:04:19 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
mariadb/mysql-test/main/vector.test
Sergei Golubchik b09c8b03d7 MDEV-35244 Vector-related system variables could use better names 
considering that users don't interact with MariaDB vector search directly,
but primarily use AI frameworks, we should use names familiar
to vector store connector writers and for AI framework users.
That is industry standard M and ef.

mhnsw_cache_size -> mhnsw_max_cache_size
mhnsw_distance_function -> mhnsw_default_distance
mhnsw_max_edges_per_node -> mhnsw_default_m
mhnsw_min_limit -> mhnsw_ef_search

inside CREATE TABLE:
max_edges_per_node -> m
distance_function -> distance
2024-11-05 14:00:52 -08:00

428 lines
16 KiB
Text
Raw Blame History

replace_result InnoDB MyISAM;
error ER_NO_INDEX_ON_TEMPORARY;
create temporary table t1 (id int auto_increment primary key, v vector(5) not null, vector index (v));
error ER_NOT_SUPPORTED_YET;
create table t1 (id int auto_increment primary key,
u vector(5) not null, vector index (u),
v vector(5) not null, vector index (v));
create table t1 (id int auto_increment primary key, v vector(5) not null, vector index (v));
replace_result InnoDB MyISAM;
show create table t1;
show keys from t1;
drop table t1;
set mhnsw_default_m=@@mhnsw_default_m+1;
create table t1 (id int auto_increment primary key, v vector(5) not null, vector index (v));
replace_result InnoDB MyISAM;
show create table t1;
show keys from t1;
drop table t1;
create table t1 (id int auto_increment primary key, v vector(5) not null,
vector index (v) m=5);
replace_result InnoDB MyISAM;
show create table t1;
show keys from t1;
set mhnsw_default_m=default;
query_vertical select * from information_schema.statistics where table_name='t1';
# print unpack("H*",pack("f*",map{rand}1..5))
insert t1 (v) values (x'e360d63ebe554f3fcdbc523f4522193f5236083d'),
(x'f511303f72224a3fdd05fe3eb22a133ffae86a3f'),
(x'f09baa3ea172763f123def3e0c7fe53e288bf33e'),
(x'b97a523f2a193e3eb4f62e3f2d23583e9dd60d3f'),
(x'f7c5df3e984b2b3e65e59d3d7376db3eac63773e'),
(x'de01453ffa486d3f10aa4d3fdd66813c71cb163f'),
(x'76edfc3e4b57243f10f8423fb158713f020bda3e'),
(x'56926c3fdf098d3e2c8c5e3d1ad4953daa9d0b3e'),
(x'7b713f3e5258323f80d1113d673b2b3f66e3583f'),
(x'6ca1d43e9df91b3fe580da3e1c247d3f147cf33e');
select id, hex(v), vec_totext(v) from t1;
flush tables;
# test with a valid query vector
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(v, x'B047263c9f87233fcfd27e3eae493e3f0329f43e') d from t1 order by d limit 3;
# swapped arguments
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(x'b047263C9f87233fcfd27e3eae493e3f0329f43e', v) d from t1 order by d limit 3;
# test with NULL (id is unpredictable)
--replace_regex /(\.\d{5})\d+/\1/
select id>0,vec_distance_euclidean(v, NULL) d from t1 order by d limit 3;
# test with invalid query vector (id is unpredictable)
--replace_regex /(\.\d{5})\d+/\1/
select id>0,vec_distance_euclidean(v, x'123456') d from t1 order by d limit 3;
--replace_regex /(\.\d{5})\d+/\1/
select t1.id as id1, t2.id as id2, vec_distance_euclidean(t1.v, t2.v) from t1, t1 as t2 order by 3,1,2;
# where clause
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(v, x'b047263C9F87233fcfd27e3eae493e3f0329f43e') d
from t1 order by d limit 9;
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(v, x'b047263C9F87233fcfd27e3eae493e3f0329f43e') d
from t1 where id % 3 = 0 order by d limit 3;
# subquery
--replace_regex /(\.\d{5})\d+/\1/
select * from (
select id,vec_distance_euclidean(v, x'b047263C9F87233fcfd27e3eae493e3f0329f43e') d
from t1 where id < 10
) u order by d limit 3;
# see if order by uses index:
--disable_view_protocol
--disable_ps2_protocol
flush session status;
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(v, x'B047263c9f87233fcfd27e3eae493e3f0329f43e') d from t1 order by d limit 3;
show status like 'handler_read_rnd_next'; # used
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(v, x'B047263c9f87233fcfd27e3eae493e3f0329f43e') d from t1 use index () order by d limit 3;
show status like 'handler_read_rnd_next'; # not used
flush session status;
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_cosine(v, x'B047263c9f87233fcfd27e3eae493e3f0329f43e') d from t1 order by d limit 3;
show status like 'handler_read_rnd_next'; # not used, wrong distance metric
--enable_ps2_protocol
--enable_view_protocol
# test delete
delete from t1 where v = x'7b713f3e5258323f80d1113d673b2b3f66e3583f';
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(v, x'B047263C9f87233fcfd27e3eae493e3f0329f43e') d from t1 order by d limit 3;
# test insert deleted vec
insert t1 (v) values (x'7b713f3e5258323f80d1113d673b2b3f66e3583f');
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(v, x'b047263c9F87233fcfd27e3eae493e3f0329f43e') d from t1 order by d limit 3;
# test update
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(v, x'B047263c9F87233fcfd27e3eae493e3f0329f43e') d from t1 order by d limit 5;
update t1 set v=x'76EDFC3E4B57243F10F8423FB158713F020BAA3E' where v=x'6CA1D43E9DF91B3FE580DA3E1C247D3F147CF33E';
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(v, x'B047263C9F87233fcfd27e3eae493e3f0329f43e') d from t1 order by d limit 5;
# test delete all and reinsert
delete from t1;
insert t1 (v) values (x'e360d63ebe554f3fcdbc523f4522193f5236083d'),
(x'f511303f72224a3fdd05fe3eb22a133ffae86a3f'),
(x'f09baa3ea172763f123def3e0c7fe53e288bf33e'),
(x'b97a523f2a193e3eb4f62e3f2d23583e9dd60d3f'),
(x'f7c5df3e984b2b3e65e59d3d7376db3eac63773e'),
(x'de01453ffa486d3f10aa4d3fdd66813c71cb163f'),
(x'76edfc3e4b57243f10f8423fb158713f020bda3e'),
(x'56926c3fdf098d3e2c8c5e3d1ad4953daa9d0b3e'),
(x'7b713f3e5258323f80d1113d673b2b3f66e3583f'),
(x'6ca1d43e9df91b3fe580da3e1c247d3f147cf33e');
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(v, x'b047263c9f87233Fcfd27e3eae493e3f0329f43e') d from t1 order by d limit 5;
--error ER_TRUNCATED_WRONG_VALUE_FOR_FIELD
insert t1 (v) values ('');
--error ER_TRUNCATED_WRONG_VALUE_FOR_FIELD
insert t1 (v) values (x'1234');
--error ER_TRUNCATED_WRONG_VALUE_FOR_FIELD
insert t1 (v) values (x'12345678');
drop table t1;
let $datadir=`select @@datadir`;
list_files $datadir/test;
--echo # Check if CREATE TABLE ... LIKE inherits VECTOR index
create table t1 (id int auto_increment primary key, v vector(5) not null, vector index (v));
create table t2 like t1;
replace_result InnoDB MyISAM;
show create table t2;
drop table t1, t2;
list_files $datadir/test;
--echo # Test insert ... select with vector index
create table t1 (id int auto_increment primary key, v vector(5) not null, vector index (v));
create table t2 like t1;
insert t1 (v) values (x'e360d63ebe554f3fcdbc523f4522193f5236083d'),
(x'f511303f72224a3fdd05fe3eb22a133ffae86a3f'),
(x'f09baa3ea172763f123def3e0c7fe53e288bf33e'),
(x'b97a523f2a193e3eb4f62e3f2d23583e9dd60d3f'),
(x'f7c5df3e984b2b3e65e59d3d7376db3eac63773e'),
(x'de01453ffa486d3f10aa4d3fdd66813c71cb163f'),
(x'76edfc3e4b57243f10f8423fb158713f020bda3e'),
(x'56926c3fdf098d3e2c8c5e3d1ad4953daa9d0b3e'),
(x'7b713f3e5258323f80d1113d673b2b3f66e3583f'),
(x'6ca1d43e9df91b3fe580da3e1c247d3f147cf33e');
insert into t2 select id+10, v from t1;
insert into t1 select * from t2;
select id, hex(v) from t1;
drop table t1, t2;
list_files $datadir/test;
create table t1 (id int auto_increment primary key, v vector(5) not null, vector index (v));
insert t1 (id, v) values (1, x'e360d63ebe554f3fcdbc523f4522193f5236083d');
truncate table t1;
if ($MTR_COMBINATION_MYISAM) {
--replace_result $datadir datadir
--exec $MYISAMCHK -d $datadir/test/t1#i#01
}
if ($MTR_COMBINATION_ARIA) {
--replace_result $datadir datadir
--exec $MARIA_CHK -d $datadir/test/t1#i#01
}
insert t1 (id, v) values (1, x'e360d63ebe554f3fcdbc523f4522193f5236083d');
truncate table t1;
if ($MTR_COMBINATION_MYISAM) {
--replace_result $datadir datadir
--exec $MYISAMCHK -d $datadir/test/t1#i#01
}
if ($MTR_COMBINATION_ARIA) {
--replace_result $datadir datadir
--exec $MARIA_CHK -d $datadir/test/t1#i#01
}
insert t1 (id, v) values (1, x'e360d63ebe554f3fcdbc523f4522193f5236083d');
select id, hex(v) from t1;
replace_result InnoDB MyISAM;
show create table t1;
drop table t1;
list_files $datadir/test;
--echo # Test RENAME TABLE with vector index
create table t1 (id int auto_increment primary key, v vector(5) not null, vector index (v));
list_files $datadir/test;
rename table t1 to t2;
list_files $datadir/test;
create database test1;
rename table test.t2 to test1.t1;
list_files $datadir/test1;
if ($MTR_COMBINATION_MYISAM) {
remove_file $datadir/test1/t1#i#01.MYD;
}
if ($MTR_COMBINATION_ARIA) {
remove_file $datadir/test1/t1#i#01.MAD;
}
if ($MTR_COMBINATION_INNODB) {
call mtr.add_suppression('InnoDB: Cannot rename.*because the source file does not exist');
call mtr.add_suppression('InnoDB: File ./test1/t1#i#01.ibd was not found');
remove_file $datadir/test1/t1#i#01.ibd;
}
--error 7,29,ER_ERROR_ON_RENAME
rename table test1.t1 to test1.t2;
list_files $datadir/test1;
disable_warnings;
drop database test1;
enable_warnings;
list_files $datadir/test;
--echo #
--echo # Cosine distance
--echo #
create table t1 (id int auto_increment primary key, v vector(5) not null,
vector index (v) distance=cosine);
replace_result InnoDB MyISAM;
show create table t1;
insert t1 (v) values (x'e360d63ebe554f3fcdbc523f4522193f5236083d'),
(x'f511303f72224a3fdd05fe3eb22a133ffae86a3f'),
(x'f09baa3ea172763f123def3e0c7fe53e288bf33e'),
(x'b97a523f2a193e3eb4f62e3f2d23583e9dd60d3f'),
(x'f7c5df3e984b2b3e65e59d3d7376db3eac63773e'),
(x'de01453ffa486d3f10aa4d3fdd66813c71cb163f'),
(x'76edfc3e4b57243f10f8423fb158713f020bda3e'),
(x'56926c3fdf098d3e2c8c5e3d1ad4953daa9d0b3e'),
(x'7b713f3e5258323f80d1113d673b2b3f66e3583f'),
(x'6ca1d43e9df91b3fe580da3e1c247d3f147cf33e');
# make sure the graph is loaded
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_cosine(v, x'B047263c9f87233fcfd27e3eae493e3f0329f43e') d from t1 order by d limit 3;
--disable_view_protocol
--disable_ps2_protocol
flush session status;
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_cosine(v, x'B047263c9f87233fcfd27e3eae493e3f0329f43e') d from t1 order by d limit 3;
show status like 'handler_read_rnd_next';
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_cosine(v, x'B047263c9f87233fcfd27e3eae493e3f0329f43e') d from t1 use index () order by d limit 3;
show status like 'handler_read_rnd_next';
flush session status;
--replace_regex /(\.\d{5})\d+/\1/
select id,vec_distance_euclidean(v, x'B047263c9f87233fcfd27e3eae493e3f0329f43e') d from t1 order by d limit 3;
show status like 'handler_read_rnd_next';
--enable_ps2_protocol
--enable_view_protocol
drop table t1;
# distance to itself
set @a=vec_fromtext('[94.542572,8.735560,60.050098,74.043800,90.068710,28.212160,70.854660,69.636841,35.620232,69.190628]');
select vec_distance_cosine(@a, @a), vec_distance_euclidean(@a, @a);
--echo # Test ALTER TABLE, CREATE/DROP INDEX
create table t1 (id int auto_increment primary key, v vector(5) not null, vector index (v));
insert t1 (v) values (x'e360d63ebe554f3fcdbc523f4522193f5236083d'),
(x'f511303f72224a3fdd05fe3eb22a133ffae86a3f'),
(x'f09baa3ea172763f123def3e0c7fe53e288bf33e'),
(x'b97a523f2a193e3eb4f62e3f2d23583e9dd60d3f'),
(x'f7c5df3e984b2b3e65e59d3d7376db3eac63773e'),
(x'de01453ffa486d3f10aa4d3fdd66813c71cb163f'),
(x'76edfc3e4b57243f10f8423fb158713f020bda3e'),
(x'56926c3fdf098d3e2c8c5e3d1ad4953daa9d0b3e'),
(x'7b713f3e5258323f80d1113d673b2b3f66e3583f'),
(x'6ca1d43e9df91b3fe580da3e1c247d3f147cf33e');
--echo # ADD/DROP COLUMN, ALGORITHM=COPY
alter table t1 add column a int, algorithm=copy;
list_files $datadir/test;
show create table t1;
alter table t1 drop column a, algorithm=copy;
list_files $datadir/test;
show create table t1;
--echo # ADD/DROP INDEX, ALGORITHM=COPY (non-vector)
alter table t1 add index a(id), algorithm=copy;
list_files $datadir/test;
show create table t1;
alter table t1 drop index a, algorithm=copy;
list_files $datadir/test;
show create table t1;
--echo # CREATE/DROP INDEX, ALGORITHM=COPY (non-vector)
create index a on t1(id) algorithm=copy;
list_files $datadir/test;
show create table t1;
drop index a on t1;
list_files $datadir/test;
show create table t1;
--echo # ADD/DROP COLUMN IF [NOT] EXISTS, ALGORITHM=COPY (non-vector)
alter table t1 add column if not exists a int, algorithm=copy;
list_files $datadir/test;
show create table t1;
alter table t1 drop column if exists a, algorithm=copy;
list_files $datadir/test;
show create table t1;
--echo # ADD/DROP INDEX, ALGORITHM=COPY (vector)
alter table t1 drop index v, algorithm=copy;
list_files $datadir/test;
show create table t1;
alter table t1 add vector index v(v), algorithm=copy;
list_files $datadir/test;
show create table t1;
--echo # CREATE/DROP INDEX, ALGORITHM=COPY (vector)
drop index v on t1;
list_files $datadir/test;
show create table t1;
create vector index v on t1(v) algorithm=copy;
list_files $datadir/test;
show create table t1;
--echo # ADD/DROP INDEX, ALGORITHM=INPLACE (non-vector)
--error ER_ALTER_OPERATION_NOT_SUPPORTED
alter table t1 add index a(id), algorithm=inplace;
alter table t1 add index a(id);
--error ER_ALTER_OPERATION_NOT_SUPPORTED
alter table t1 drop index a, algorithm=inplace;
alter table t1 drop index a;
--echo # ADD/DROP INDEX, ALGORITHM=INPLACE (vector)
--error ER_ALTER_OPERATION_NOT_SUPPORTED
alter table t1 drop index v, algorithm=inplace;
alter table t1 drop index v;
--error ER_ALTER_OPERATION_NOT_SUPPORTED
alter table t1 add vector index v(v), algorithm=inplace;
alter table t1 add vector index v(v);
--echo # CHANGE/DROP/MODIFY COLUMN, ALGORITHM=INPLACE (vector)
--error ER_ALTER_OPERATION_NOT_SUPPORTED
alter table t1 change column v v vector(6) not null, algorithm=inplace;
--error ER_ALTER_OPERATION_NOT_SUPPORTED
alter table t1 drop column v, algorithm=inplace;
--error ER_ALTER_OPERATION_NOT_SUPPORTED
alter table t1 modify column v vector(7) not null, algorithm=inplace;
--echo # ADD/CHANGE/DROP/MODIFY COLUMN, ALGORITHM=INPLACE (non-vector)
if ($MTR_COMBINATION_INNODB) {
alter table t1 add column a varchar(10), algorithm=inplace;
alter table t1 change column a a varchar(20), algorithm=inplace;
alter table t1 modify column a varchar(30), algorithm=inplace;
alter table t1 drop column a, algorithm=inplace;
list_files $datadir/test;
show create table t1;
}
--echo # ENABLE/DISABLE INDEXES
alter table t1 disable keys;
alter table t1 enable keys;
list_files $datadir/test;
show create table t1;
--echo # RENAME COLUMN (vector)
alter table t1 rename column v to w;
list_files $datadir/test;
show create table t1;
alter table t1 rename column w to v;
--echo # RENAME INDEX (vector)
alter table t1 rename key v to w;
list_files $datadir/test;
show create table t1;
alter table t1 rename key w to v;
--echo # IF [NOT] EXISTS
create vector index if not exists v on t1(v);
drop index if exists v on t1;
drop index if exists v on t1;
list_files $datadir/test;
show create table t1;
create vector index if not exists v on t1(v);
alter table t1 rename key if exists v to w;
alter table t1 rename key if exists w to v;
alter table t1 alter key if exists v ignored;
alter table t1 alter key if exists v not ignored;
--echo # ENGINE
if ($MTR_COMBINATION_INNODB) {
alter table t1 engine=myisam;
list_files $datadir/test;
show create table t1;
alter table t1 engine=innodb;
list_files $datadir/test;
show create table t1;
alter table t1 rename to t2, engine=myisam;
list_files $datadir/test;
show create table t2;
alter table t2 rename to t1, engine=innodb;
list_files $datadir/test;
show create table t1;
}
--echo # CHANGE/MODIFY/DROP COLUMN (vector)
--error ER_WRONG_ARGUMENTS
alter table t1 modify column v int;
--error ER_WRONG_ARGUMENTS
alter table t1 change column v v int;
--error ER_INDEX_CANNOT_HAVE_NULL
alter table t1 modify column v vector(5);
--error ER_INDEX_CANNOT_HAVE_NULL
alter table t1 change column v v vector(6);
alter table t1 modify column v vector(7) not null;
list_files $datadir/test;
show create table t1;
alter table t1 change column v v vector(5) not null;
list_files $datadir/test;
show create table t1;
alter table t1 drop column v;
list_files $datadir/test;
show create table t1;
drop table t1;
# This is supposed to test crash while filling indexes_option_struct array,
# which doesn't happen because alloc_root(0) returns something. Add a test
# anyway and fix indexes_option_struct array allocation.
create table t1(v vector(5) not null, vector index(v));
alter table t1 add column a int;
drop table t1;
Reference in a new issue View git blame Copy permalink