Heikki Linnakangas wrote:
> Ned T. Crigler wrote:
>> It appears that the visibility map patch is causing pg_class.reltuples
>> to be
>> set improperly after a vacuum. For example, it is set to 0 if the map
>> indicated that no pages in the heap needed to be scanned.
>>
>> Perhaps reltuples should not be updated unless every page was scanned
>> during
>> the vacuum?
>
> Yeah, vacuum shouldn't overwrite reltuples if it hasn't scanned all pages.
Because we use reltuples divided by relpages in the planner, we probably
shouldn't update relpages either if we don't update reltuples.
Otherwise, if the table has grown a lot since we last updated reltuples,
the reltuples / relpages ratio would be less, not more, accurate, if
relpages is updated to a new higher value but reltuples is not.
-- Heikki Linnakangas EnterpriseDB http://www.enterprisedb.com