I'm quite new to Postgres so I don't know how to read the execution
plans. However it is obvious that the plans for the query with and
without the function are different. I added some indices to the table
roleassertions and it seems to solve this big difference in the
execution times of both queries. So I guess it was not the function that
was the problem in the end. Thanks a lot!
Inanc
On 5/11/12 5:55 PM, Tom Lane wrote:
> Inanc Seylan<inanc.seylan@gmail.com> writes:
>> It is IMMUTABLE. I attach the output of EXPLAIN both with and without
>> the simple function (returning true only) in the query.
>
> EXPLAIN ANALYZE would have been far more helpful. However, the thing
> that jumps out at me here is all the seqscans on table "symbols".
> Do you not have an index on symbols.id? If you do, perhaps there is
> a datatype-mismatch problem preventing it from being used.
>
> regards, tom lane