Stephan Szabo <sszabo@megazone23.bigpanda.com> writes:
> Yes, column = NULL should *never* return true according to the spec (it
> should always return NULL in fact as stated). The reason for breaking
> with the spec is AFAIK to work with broken microsoft clients that seem to
> think that =NULL is a meaningful test and generate queries using that.
Microsoft Access is the guilty party, IIRC. I recently tried to stir up
some interest in changing this behavior back to the standard, but
apparently there are still too many people using broken versions of
Access.
A compromise answer might be to offer a SET variable that selects the
Microsoft-compatible misimplementation. Would that fly?
regards, tom lane