Re: A first table partitioning exercise.

[Mattias Jonsson]

I cannot repeat your minor difference. The worst case I got for the same test was 1:9. I have tested both InnoDB and MyISAM with the same results, using the latest source built with release configuration, and started through './mtr --start partition' in the mysql-test directory.

Which build are you using? is it a debug build? on my 3-years old macbook the times are:
mysql> select count(*) from no_part_tab where c3 > '1995-01-01' and c3 < '1995-12-31';
+----------+
| count(*) |
+----------+
| 795181 |
+----------+
1 row in set (3.05 sec)

mysql> select count(*) from part_tab where c3 > '1995-01-01' and c3 < '1995-12-31';
+----------+
| count(*) |
+----------+
| 795181 |
+----------+
1 row in set (0.33 sec)

mysql> select count(*) from i_part where c3 > '1995-01-01' and c3 < '1995-12-31';
+----------+
| count(*) |
+----------+
| 795181 |
+----------+
1 row in set (0.71 sec)

mysql> select count(*) from i_no_part where c3 > '1995-01-01' and c3 < '1995-12-31';
+----------+
| count(*) |
+----------+
| 795181 |
+----------+
1 row in set (7.02 sec)

No real difference between latest 5.1 and 5.5 (i_* tables are InnoDB).

If you have a repeatable test setup, could you please add that here?