Re: Fooling the query optimizer
От | Adam Lang |
---|---|
Тема | Re: Fooling the query optimizer |
Дата | |
Msg-id | 013a01c091eb$d4e20c20$330a0a0a@rutgersinsurance.com обсуждение исходный текст |
Ответ на | Fooling the query optimizer ("Brent R.Matzelle" <bmatzelle@yahoo.com>) |
Список | pgsql-php |
Let me look into this. Adam Lang Systems Engineer Rutgers Casualty Insurance Company http://www.rutgersinsurance.com ----- Original Message ----- From: "Brent R.Matzelle" <bmatzelle@yahoo.com> To: "PostgreSQL PHP" <pgsql-php@postgresql.org> Sent: Thursday, February 08, 2001 10:41 AM Subject: [PHP] Fooling the query optimizer > Have any of you discovered a way to get around the current query optimizer > limitation in Postgres? For example, I have a table that has three columns > that I want to index for frequent search duties. In Postgres I am forced to > create three indicies: one including all three columns, one for col2 and > col3, and one for just col3. Databases like MySQL can use the first index > for these types of queries "SELECT * WHERE col2 = x AND col3 = y" and "SELECT > * WHERE col3 = y". Postgres could only perform queries on indicies where it > looks like "SELECT * WHERE col1 = x AND col2 = y AND col3 = z" and "SELECT * > WHERE col1 = x AND col2 = y" etc. However adding extra indexes as above > would decrease the write speed on that table because a simple insert would > require an update on all three indicies. > > Is there a way to fool Postgres to use the first index by creating a query > like "SELECT * WHERE col1 = * AND col3 = x"? I know I'm grasping for straws > here, but these issues can kill my database query performance. > > Brent
В списке pgsql-php по дате отправления: