Details
-
Bug
-
Status: Stalled (View Workflow)
-
Major
-
Resolution: Unresolved
-
10.2(EOL), 10.3(EOL), 10.4(EOL)
Description
--source include/have_rocksdb.inc
Dataset
create table t1 (id1 bigint, id2 bigint, c1 bigint, c2 bigint, c3 bigint, c4 bigint, c5 bigint, c6 bigint, c7 bigint, primary key (id1, id2), index i(c1, c2))engine=rocksdb; |
 |
let $i=0;
|
while ($i<10000)
|
{
|
inc $i;
|
eval insert t1(id1, id2, c1, c2, c3, c4, c5, c6, c7) |
values($i, 0, $i, 0, 0, 0, 0, 0, 0); |
}
|
explain select c1 from t1 where c1 > 5 limit 10; |
id select_type table type possible_keys key key_len ref rows Extra
|
1 SIMPLE t1 range i i 9 NULL 13300 Using where; Using index
|
If you look closely the optimizer predicts reading 13300 rows but the table t1 has in total 10000 rows. These estimates look incorrect.