Stef <svb(at)ucs(dot)co(dot)za> writes:
> Reducing sort_mem to 4096 seems to make it run in a reasonable time
> again. Any idea why? The database does a whole lot of huge sorts
> every day, so I thought upping this parameter would help.
Not if you haven't got the RAM to support it :-(
Another thing you might look at is ANALYZEing the tables again after
you've loaded all the new data. The row-count estimates seem way off
in these plans. You might need to increase the statistics target,
too, to get better plans.
regards, tom lane