Re: Why is seq search preferred here by planner?

Mallah,

> > The reason that the planner is using a seq scan on
personal_account_details is the same as the
> > reason for using a seq scan on email_bank; the number of rows which
match the condition,
> > about 150,000. With that many qualifying rows, a seq scan is faster.
>
>
> But there are two tables here , email_bank and personal_account_details
> in personal account details only one row is supposed to match a given userid
> as userid is the PKEY , why seq_scan there ? or am i getting the explain
wrong ?

It doesn't matter whether it's a primary key or not. If your query or
subquery condition returns a large number/proportion of rows, a seq scan is
going to be faster than an index scan.

As far as userid being a key, this affects the planner's *join method*, *not*
how the database gets the rows.

--
-Josh Berkus
Aglio Database Solutions
San Francisco