Klint Gore <kg@xxxxxxxxxxxxxx> writes: > Is there any way to create operators to point like to ilike? There > doesn't seem to be a like or ilike in pg_operator (not in 7.4 anyway). Actually it's the other way 'round: if you look into gram.y you'll see that LIKE is expanded as the operator ~~ and ILIKE as the operator ~~* ... so one of the alternatives I was thinking of offering to Paul was to rename those two operators to swap 'em. However I'm afraid that that would break the planner, which has some hardwired assumptions about the behavior of those two operator OIDs. Maybe we should change the planner to look a level deeper and see what functions the operators refer to. regards, tom lane
