[koxuz]

hi,

is it normal that when i issue a "select count(*) from <table>" command postgres performs a sequential scan in order to evaluate the number of items in that table? i have a table with almost 3 million entires and the operations takes several seconds to complete. the server has plenty of ram and is fast enough.

koxuz

[eperich]

this is a well known problem in postgresql

this has to be I think of some transactional reasons

but there are attempts to getting rid of this seq scan

[karim2k]

In order to to get fatsre reponse you could shorten you query by just counting one row select count(row) from <table>, or just get the max of the id if u have a regular counter , or just create a row updated with a trigger (on upsdte/insert/delete) that saves the number of rows.

Good Luck !