I see in your 'Explain's, the 2 queries use different indexes, the fast
one uses soc_date_idx, and the slower one uses q_idx. The trick,
perhaps is to force
soc_date_idx to be used in the 2nd case. (Adding ORDER BY soc_date might
do it, 'soc_date=X and (queue_id=Y and server_id=Z) ) might do
Hi,
Can anyone please explain the following?
I do two select count(*)'s from a big table, (21,000,000 + rows 6.5 Gb).
All columns in the where are indexed. The first select is very quick, the
second very slow.
The ONLY difference is the value of one column variable (queue_id), which is
an