Re: Add RESPECT/IGNORE NULLS and FROM FIRST/LAST options

Oliver Ford <ojford@gmail.com>

From: Oliver Ford <ojford@gmail.com>

To: Tatsuo Ishii <ishii@postgresql.org>

Cc: krasiyan@gmail.com, tgl@sss.pgh.pa.us, vik@postgresfriends.org, pgsql-hackers@postgresql.org, andrew@tao11.riddles.org.uk, david@fetter.org

Date: 2025-03-06T09:57:30Z

Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

Fix coding style with "else".
- 20628b62e46e 19 (unreleased) landed
Fix multi WinGetFuncArgInFrame/Partition calls with IGNORE NULLS.
- 2d7b247cb414 19 (unreleased) landed
Fix Coverity issue reported in commit 2273fa32bce.
- dd766a441d69 19 (unreleased) landed
Use ereport rather than elog in WinCheckAndInitializeNullTreatment.
- 5f3808646f67 19 (unreleased) landed
Avoid uninitialized-variable warnings from older compilers.
- 71540dcdcb22 19 (unreleased) cited
Fix Coverity issues reported in commit 25a30bbd423.
- 2273fa32bce7 19 (unreleased) landed
Improve EXPLAIN's display of window functions.
- 8b1b342544b6 18.0 cited
Automatically generate node support functions
- 964d01ae90c3 16.0 cited

Attachments

0009-ignore-nulls.patch (application/octet-stream) patch 0009

On Fri, Feb 28, 2025 at 11:49 AM Tatsuo Ishii <ishii@postgresql.org> wrote:

> >> BTW, I noticed that in the code path where
> >> ignorenulls_getfuncarginframe() is called, WinSetMarkPosition() is
> >> never called?
> >>
> >> Attached version uses the mark_pos at the end.
>
> I did simple performance test against v8.
>
> EXPLAIN ANALYZE
> SELECT
>   x,
> nth_value(x,2) IGNORE NULLS OVER w
> FROM generate_series(1,$i) g(x)
> WINDOW w AS (ORDER BY x ROWS BETWEEN 2 PRECEDING AND 2 FOLLOWING);
>
> I changed $i = 1k, 2k, 3k, 4k, 5k... 10k and got this:
>
> Number  Time (ms)
> of rows
> ----------------
> 1000    28.977
> 2000    96.556
> 3000    212.019
> 4000    383.615
> 5000    587.05
> 6000    843.23
> 7000    1196.177
> 8000    1508.52
> 9000    1920.593
> 10000   2514.069
>
> As you can see, when the number of rows = 1k, it took 28 ms. For 10k
> rows, it took 2514 ms, which is 86 times slower than the 1k case. Can
> we enhance this?
>
>
Attached version removes the non-nulls array. That seems to speed
everything up.  Running the above query with 1 million rows averages 450ms,
similar when using lead/lag.