I have a table (let's call it
extra_info) in MySQL that has various extra information/labels on user accounts:
account_id / data_key 11 / 'test' 11 / 'blah' 12 / 'coupon a' 12 / 'coupon b' 15 / 'whatever' 18 / 'test'
And so on.
I am building reports on other tables but need to filter out records that are present in the extra_info table as 'test'.
So in the example above, records 1-10, 13-14, 16-17 are fine because they are not in extra_info, 11 needs to be filtered out (it is 'test'), 12 is fine (but obviously just need to have it once even though is has 2 labels, 15 is fine (not 'test'), and 18 needs to be removed.
I have done this with subquery in the
WHERE -- filter out test labels a.id NOT IN (SELECT ei.account_id FROM extra_info AS ei WHERE ei.data_key = 'test')
This does its job, however, I have a suspicion that a regular join would have worked faster. What's the right syntax for it?
If I just join and filter out the ones with data_key='test', #11 would still be there - so this is not a solution. Grouping by account_id will probably has performance implications and basically same impact as the current code, right?