Re: BUG #8598: Row count estimates of partial indexes
| От | Tom Lane |
|---|---|
| Тема | Re: BUG #8598: Row count estimates of partial indexes |
| Дата | |
| Msg-id | 28474.1384705755@sss.pgh.pa.us обсуждение исходный текст |
| Ответ на | BUG #8598: Row count estimates of partial indexes (marko@joh.to) |
| Ответы |
Re: BUG #8598: Row count estimates of partial indexes
|
| Список | pgsql-bugs |
marko@joh.to writes:
> We have the following partial index on a small subset of a larger table:
> "index_transactions_transaction_balance_details" btree (transactionid)
> WHERE NOT processed AND accountbalancesdailyid IS NOT NULL
> However, querying with the WHERE clause completely ignores the
> pg_class.reltuples value for the index:
Yup. Row count estimates are derived by estimating the selectivity of the
given WHERE clauses and multiplying by the (estimated) current table size.
In the particular case you show here, with a partial index that *exactly*
matches the WHERE clause, we could get a better answer by looking at the
index size --- but that doesn't scale to any less simplistic case, such
as a query with additional WHERE clauses.
It's also important to realize that reltuples for an index is a whole lot
less trustworthy than it is for a table; ANALYZE doesn't update the
former, for example. And scaling from the last-reported VACUUM stats
to current reality is going to be shakier.
So on the whole, I don't think this would be a good idea.
regards, tom lane
В списке pgsql-bugs по дате отправления: