Comment 4 for bug 999251

The very same execution plan was chosen for ANY run of Q13 at SF30. This includes default (volatile) statistics, persistent (but randomly gathered) statistics and exact statistics from MariaDB-MWL248.

I.e. here is the plan from MariaDB-5.5.23-MWL248

+------+-------------+------------+------+---------------+-------------+---------+-------------------------+---------+----------+---------------------------------------------------------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | filtered | Extra |
+------+-------------+------------+------+---------------+-------------+---------+-------------------------+---------+----------+---------------------------------------------------------------------------------+
| 1 | PRIMARY | <derived2> | ALL | NULL | NULL | NULL | NULL | 0 | 0.00 | Using temporary; Using filesort |
| 2 | DERIVED | customer | ALL | NULL | NULL | NULL | NULL | 4500000 | 100.00 | Using temporary; Using filesort |
| 2 | DERIVED | orders | ref | i_o_custkey | i_o_custkey | 5 | dbt3.customer.c_custkey | 15 | 100.00 | Using where; Using join buffer (flat, BKA join); Key-ordered Rowid-ordered scan |
+------+-------------+------------+------+---------------+-------------+---------+-------------------------+---------+----------+---------------------------------------------------------------------------------+

Execution plans for all 3 kinds of statistics are archived in lp:~ahel/maria/mariadb-benchmarks in the dbt3-runs/innodb_s30_test_2012-04 subdirectory.