Robert martin wrote:

> 687 fetches, 10 marks, 0 reads, 10 writes.
> 0 inserts, 0 updates, 0 deletes, 186 index, 0 seq.
> Delta memory: -2396 bytes.
> Execute time: 00:00:00.
>
> I say that because the above is a simplified (improved) version of an
> SQL that was run immediately prior that had the following results.
>
> 35445 fetches, 30 marks, 0 reads, 30 writes.
> 0 inserts, 0 updates, 0 deletes, 860 index, 15899 seq.
> Delta memory: 44880 bytes.
> Execute time: 00:00:00.

Obviously the first query allocated 'fresh' memory, so the second one
used that. And not entirely, since it released some as well.

In any case, the second one has zero 'sequential' reads, and even less
index reads, so if they produce the same results, that one is much
better (just as you wrote).

--
Milan Babuskov
http://www.flamerobin.org
http://www.guacosoft.com