Chris Ruprecht <chrup999@yahoo.com> writes:
> Thank you so very much for your research! I have broken down the load
> procedure into two parts and it's looking great. The run time is sort of
> identical (about 3 minutes for today's file) but memory usage stays under 30
> MB instead of rising up to ... The old routine is busy running to compare
> and it currently sits at 344 MB - major difference (377 MB now). I see a
> very slight increase in the 'size' field (under top), about 128 KB every 4
> seconds which, maybe, could be attributed to some procedural overhead?
Residual leakage within the outer procedure's for-loop, no doubt.
As long as you can live with it until 7.2 comes out, I think you'll be
okay.
regards, tom lane