On 3/9/03 3:21 PM, Tim Bunce wrote: > For postgres the "<tablename>_<columnname>_seq" name is part of the core > system so there's really no "guessing" going on (as I understand it).
That's only true for the SERIAL pseudo-type's auto-generated sequences. A table could have its primary key column default to the next value from any sequence under any name, in which case the DBD would really have no way of knowing which sequence contains the "last insert id" (unless it wants to get funky with Postgres's table introspection abilities, I guess). Anyway, I'd be happy even if only the common case of a SERIAL primary key was handled, but I don't see a good general-purpose solution to the problem. -John