diff options
| author | Monty <monty@mariadb.org> | 2022-09-23 14:48:13 +0300 |
|---|---|---|
| committer | Sergei Petrunia <sergey@mariadb.com> | 2023-02-02 23:54:57 +0300 |
| commit | 009db2288ba8d22c5e8e5e047d8f8694f7097923 (patch) | |
| tree | 96fba5cb7598afd162f464adb522b59c1ef90605 /sql/ha_partition.cc | |
| parent | b66cdbd1eaeed7e96317a03a190c496fd062ec71 (diff) | |
| download | mariadb-git-009db2288ba8d22c5e8e5e047d8f8694f7097923.tar.gz | |
Fixed limit optimization in range optimizer
The issue was that when limit is used,
SQL_SELECT::test_quick_select would set the cost of table scan to be
unreasonable high to force a range to be used.
The problem with this approach was that range was used even when the
cost of range, when it would only read 'limit rows' would be higher
than the cost of a table scan.
This patch fixes it by not accepting ranges when the range can never
have a lower cost than a table scan, even if every row would match the
WHERE clause.
Diffstat (limited to 'sql/ha_partition.cc')
| -rw-r--r-- | sql/ha_partition.cc | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/sql/ha_partition.cc b/sql/ha_partition.cc index 701bffd8dd7..30c1832fadc 100644 --- a/sql/ha_partition.cc +++ b/sql/ha_partition.cc @@ -6575,7 +6575,7 @@ ha_rows ha_partition::multi_range_read_info_const(uint keyno, RANGE_SEQ_IF *seq, void *seq_init_param, uint n_ranges, uint *bufsz, - uint *mrr_mode, + uint *mrr_mode, ha_rows limit, Cost_estimate *cost) { int error; @@ -6636,7 +6636,7 @@ ha_rows ha_partition::multi_range_read_info_const(uint keyno, &m_partition_part_key_multi_range_hld[i], m_part_mrr_range_length[i], &m_mrr_buffer_size[i], - &tmp_mrr_mode, &part_cost); + &tmp_mrr_mode, limit, &part_cost); if (tmp_rows == HA_POS_ERROR) { m_part_spec= save_part_spec; |