On 3/2/16, drum.lucas@gmail.com <drum.lucas@gmail.com> wrote:
> Hi all...
>
> I'm working on a Slow Query. It's faster now (It was 20sec before) but
> still not good.
>
> Can you have a look and see if you can find something?
> Cheers
>
> Query:
>
> WITH jobs AS (
> ...
> FROM
> jobs AS job
> JOIN
> public.ja_notes AS note
> ON
> note.jobid = job.id
> AND note.note_type IN ('time', 'part')
> ...
It is the most long part. All query is 8.8sec.
SeqScan by CTE is 2.8sec! and index scan in ix_notes_jobid_per_type
500rows(loops) * 9.878ms!!! = 4.939sec.
Why does it take so long time?
For example, index scan in ja_customers_pkey is only 0.781 per row...
10 times faster!
What definition of the ix_notes_jobid_per_type? Is it bloated?
> Explain analyze link: http://explain.depesz.com/s/IIDj
>
--
Best regards,
Vitaly Burovoy