Re: Query Tuning for NOT EQUAL TO <> operator

[Rick James]

My bad. You are not doing a JOIN.

SELECT uid, ip FROM log_count WHERE ip = '1851220710' AND uid <> '4796596'
leads to log_count needing
INDEX(ip)
and so does log_count_arch:
SELECT uid, ip FROM log_count_arch WHERE ip = '1851220710' AND uid <> '4796596' GROUP BY uid;

For log_count_arch, it might be a little more optimal to have
INDEX(ip, uid)
This might help with the GROUP BY. (It will not help any for the 'uid<>...')

EXPLAIN's rows 2 & 3 have
type: ALL
This implies doing a table scan (costly). With the suggested indexes, this is likely to change, and be faster.

Back to the title... "<>" is essentially un-optimizable.