Bruce Momjian <pgman@candle.pha.pa.us> writes:
> Tom, since you check pg_statsistic in the optimizer for the most common
> value, should we downgrade the pg_attribute.attdisbursion value when we
> know the value is not the most common value? Are you doing that
> already?
Yeah, we already do. In fact the last tweak in that area was to reduce
the ratio we multiply by in just that situation. It's a poor substitute
for actually having better stats of course...
regards, tom lane