Re: [HACKERS] 6.5 beta and ORDER BY patch

Hannu Krosing wrote:

>
> Jan Wieck wrote:
> >
> > >
> > > Hi PostgreSQL hackers
> > >
> > > As we are again approaching the beta (feature freeze),
> > > I will ask my ordinary question ;)
> > >
> > > Is the patch by Jan that eliminated the duplicate sort node in case it
> > > was redundant included in 6.5 ?
> >
> >     Sorry,
> >
> >     I  missed  to  put  it  into after v6.4 release. And since it
> >     wasn't there during v6.5 development, I would not put  it  in
> >     now.
> >
> >     Note that it wasn't in the v6.4 feature patches either, so it
> >     isn't tested enough to get released.
>
> But if it is not relesed it will _never_ be tested enough ...
>
> As we are just going into beta, not relese, I would suggest to put
> it in now, and back out if it relly breaks anything.
>
> I have been using it with 6.4 almost since the relese an have
> seen no problems - in fact it solved a big problem and provided about
> 1000X speedup for certain queries (a fraction of second instead of
> 6 minutes) , not to mention avoiding backend crashes due to disk space
> exhaustion.
>
> And it did not break anything in regression tests either, the only
> argument then was that there is nothing in regression tests that
> could possibly be broken by it ;)
>
> I greatly prefer it over my previous method of doing the same on the
> client side (issuing an EXPLAIN, parsing it to see if it is SORT on
> INDEX SCAN, and omitting the ORDER BY if it is)
>
> Also, not having it greatly diminishes the value of LIMIT.

    Ok ok ok - OK. You got me, I'll go ahead and put it in.

>
> I agree that it is a hack and only a partial solution and that in
> ideal world the optimiser would also know about sort nodes.

    First  the  executor  must  know better how to handle LIMIT's
    OFFSET.  For now it  processes  the  query  until  OFFSET  is
    reached,  simply  suppressing  the  in  fact  produced result
    tuples in the output. The it stops sending if the LIMIT count
    is  reached.   For  joins  or other complex things, it has no
    chance to do something different. But for an  indexed  single
    table  scan,  where  ALL  the  qualifications are done on the
    index, it should handle the OFFSET by skipping  index  tuples
    only.

    Second  the  optimizer  must  take  LIMIT  into  account  and
    depending on the known number of  tuples,  LIMIT  and  OFFSET
    produce  an  index  scan even if the query isn't qualified at
    all but has an ORDER BY clause matched by the index.

    These two  features  would  finally  solve  your  huge  table
    problems.


Jan

--

#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me.                                  #
#======================================== jwieck@debis.com (Jan Wieck) #