15 Aug
2008
15 Aug
'08
7:08 a.m.
Niels Nes wrote:
I have built the August 14 super tar ball, loaded the table, made sure all queries ran ok, restarted the server, then ran them both again with the -G switch. Unfortunately the total usec in the trace file are the same as before for both queries.
Using your create statements and queries I could see that the plans for both q1 and q2 are equal, ie they should be the same speed. If not something is wrong, ie wrong version or so. Question remaining is are the queries both on the same 'too' slow speed or are the fast.
How about the compiler optimizer settings? e.g. --enable-optimize and friends