> test=> delete from cash_journal where ARRAY[click, cash_journal_id] NOT in
> (select max(ARRAY[click,cash_journal_id]) from cash_journal group by
> fairian_id); DELETE 7
For what it's worth, we've run into *severe* performance issues using in() if
there are a large number of values in conjunction with a complex query. (EG:
more than 10,000) Using a with() prefix table and joining against that doesn't
seem to carry anything like a similar performance penalty.