Re: Implement targetlist SRFs using ROWS FROM() (was Changed SRF in targetlist handling)

Heikki Linnakangas <hlinnaka@iki.fi>

From: Heikki Linnakangas <hlinnaka@iki.fi>
To: Andres Freund <andres@anarazel.de>, pgsql-hackers@postgresql.org, Tom Lane <tgl@sss.pgh.pa.us>
Date: 2016-08-29T09:56:25Z
Lists: pgsql-hackers
On 08/28/2016 12:48 AM, Andres Freund wrote:
> Attached is a significantly updated patch series (see the mail one up
> for details about what this is, I don't want to quote it in its
> entirety).
>
> There's still some corner cases (DISTINCT + SRF, UNION/INTERSECT with
> SRF) to test / implement and a good bit of code cleanup to do. But
> feature wise it's pretty much complete.

Looks good, aside from the few FIXMEs, TODOs and XXXs and DIRTYs.

I think we need to come up with a better word for "unsrfify". That's 
quite a mouthful. Perhaps something as boring as 
"convert_srfs_to_function_rtes".

Would it make sense for addRangeTableEntryForFunction() to take  a List 
of RangeFunctionElems as argument, now that we have such a struct? The 
lists-of-same-length approach gets a bit tedious.

Typos:
s/fortfour/forfour
s/Each element of this list a/ Each element of this list is a/

- Heikki



Commits

  1. Remove obsoleted code relating to targetlist SRF evaluation.

  2. Doc: improve documentation of new SRF-in-tlist behavior.

  3. Move targetlist SRF handling from expression evaluation to new executor node.

  4. Don't split up SRFs when choosing to postpone SELECT output expressions.