Details
-
Bug
-
Status: Open (View Workflow)
-
Major
-
Resolution: Unresolved
-
N/A
-
None
Description
Queries which use vector indexes are written into the slow log under filter not_using_index, and the record claims they are using full scan.
Oddly, they are nevertheless not written under full_scan filter (it is not in the test case, but can be seen by changing the filter).
The test case is for reproducing purposes only, please don't put it into the regression suite as is, since it prints a raw chunk from the slow log which is bound to be non-deterministic.
--source include/have_sequence.inc
|
|
--let $logfile= `select @@slow_query_log_file`
|
|
create table t (v vector(1) not null, vector(v)); |
insert into t select vec_fromtext(concat('[',seq,']')) FROM seq_1_to_100; |
|
set log_slow_verbosity=full, log_slow_filter='not_using_index'; |
select vec_totext(v) from t order by vec_distance_euclidean(v,vec_fromtext('[0]')) limit 3; |
|
--echo ### slow log ###
|
--exec tail -n 14 $logfile
|
--echo ################
|
|
# Cleanup
|
drop table t; |
7aab5e35d714fc3b6de68f5158072d5ecdf62faa |
### slow log ###
|
# Thread_id: 4 Schema: test QC_hit: No
|
# Query_time: 0.000351 Lock_time: 0.000098 Rows_sent: 3 Rows_examined: 3
|
# Rows_affected: 0 Bytes_sent: 88
|
# Pages_accessed: 0 Pages_read: 0 Pages_prefetched: 0 Pages_updated: 0 Old_rows_read: 0
|
# Pages_read_time: 0.0000 Engine_time: 0.0071
|
# Full_scan: Yes Full_join: No Tmp_table: No Tmp_table_on_disk: No
|
# Filesort: No Filesort_on_disk: No Merge_passes: 0 Priority_queue: No
|
#
|
# explain: id select_type table type possible_keys key key_len ref rows r_rows filtered r_filtered Extra
|
# explain: 1 SIMPLE t index NULL v 6 NULL 100 3.00 100.00 100.00
|
#
|
use `test`;
|
SET timestamp=1730395492;
|
select vec_totext(v) from t order by vec_distance_euclidean(v,vec_fromtext('[0]')) limit 3;
|
################
|
Attachments
Issue Links
- is caused by
-
MDEV-34939 vector search in 11.7
- Closed