Re: Seq Scan because of stats or because of cast?

OK, I've added support for native OIDs values (i.e. unsigned int, sizeof(4)) in my libpq wrapper.

Tested with binary binding and getting of scalar and array values.

But to truly test this is working OK, I'd need OIDs in the range [2^31, 2^32),

while the OIDs in my DB only reach in the 200M range.

So, any way to force the DB to create OIDs in that special range?

Without hosing my DB / Cluster that is... This is not a throw-away DB / Cluster.

Thanks, --DD