Thread: Why does make_noname insert a SeqScan above sort/material node?

Does anyone know why make_noname (in createplan.c) insists on putting
a SeqScan plan node above the Sort or Material node it's generating?
As far as I can tell, it's a waste of cycles:

1. planner.c doesn't bother with a SeqScan above the Sorts it makes.
2. The executor's nodeSeqscan.c just redirects all its calls to the  outerPlan node, if it has an outerPlan.
3. Things seem to work fine without it ;-)

However, I'm not quite ready to commit this change without consultation.
Does anyone know what this was for?
        regards, tom lane

> Does anyone know why make_noname (in createplan.c) insists on putting
> a SeqScan plan node above the Sort or Material node it's generating?
> As far as I can tell, it's a waste of cycles:
> 
> 1. planner.c doesn't bother with a SeqScan above the Sorts it makes.
> 2. The executor's nodeSeqscan.c just redirects all its calls to the
>    outerPlan node, if it has an outerPlan.
> 3. Things seem to work fine without it ;-)
> 
> However, I'm not quite ready to commit this change without consultation.
> Does anyone know what this was for?

Remove it.  Noname is for an internal temp table and they probably used
it for some other uses in the past.  That code needs cleaning, I bet.

--  Bruce Momjian                        |  http://www.op.net/~candle maillist@candle.pha.pa.us            |  (610)
853-3000+  If your life is a hard drive,     |  830 Blythe Avenue +  Christ can be your backup.        |  Drexel Hill,
Pennsylvania19026