Horton

Re: [PATCH] Incremental sort (was: PoC: Partial sort)

James Coleman <jtc331@gmail.com>

View thread View on postgr.es
From: James Coleman <jtc331@gmail.com>
To: Tomas Vondra <tomas.vondra@2ndquadrant.com>
Cc: Michael Paquier <michael@paquier.xyz>, Rafia Sabih <rafia.pghackers@gmail.com>, Peter Geoghegan <pg@bowt.ie>, Simon Riggs <simon@2ndquadrant.com>, Shaun Thomas <shaun.thomas@2ndquadrant.com>, Dmitry Dolgov <9erthalion6@gmail.com>, Alexander Korotkov <a.korotkov@postgrespro.ru>, Tom Lane <tgl@sss.pgh.pa.us>, PostgreSQL Developers <pgsql-hackers@lists.postgresql.org>, Alvaro Herrera <alvherre@2ndquadrant.com>
Date: 2020-03-13T17:16:44Z
Lists: pgsql-hackers

Attachments

On Tue, Mar 10, 2020 at 10:44 PM Tomas Vondra
<tomas.vondra@2ndquadrant.com> wrote:
...
> Now, a couple comments about parts 0001 - 0003 of the patch ...
>
> 1) I see a bunch of failures in the regression test, due to minor
> differences in the explain output. All the differences are about minor
> changes in memory usage, like this:
>
> -               "Sort Space Used": 30,                             +
> +               "Sort Space Used": 29,                             +
>
> I'm not sure if it happens on my machine only, but maybe the test is not
> entirely stable.

make check passes on multiple machines for me; what arch/distro are you using?

Is there a better way to test these? I would prefer these code paths
have test coverage, but the standard SQL tests don't leave a good way
to handle stuff like this.

Is TAP the only alternative, and do you think it'd be worth considering?

> 2) I think this bit in ExecReScanIncrementalSort is wrong:
>
>      node->sort_Done = false;
>      tuplesort_end(node->fullsort_state);
>      node->prefixsort_state = NULL;
>      tuplesort_end(node->fullsort_state);
>      node->prefixsort_state = NULL;
>      node->bound_Done = 0;
>
> Notice both places reset fullsort_state and set prefixsort_state to
> NULL. Another thing is that I'm not sure it's fine to pass NULL to
> tuplesort_end (my guess is tuplesort_free will fail when it gets NULL).

Fixed.

James

Commits

  1. Further adjustments to Hashagg EXPLAIN ANALYZE output

  2. Rework EXPLAIN format for incremental sort

  3. Fix typos and improve incremental sort comments

  4. Stabilize incremental_sort tests

  5. Minor improvements in Incremental Sort explain

  6. Consider Incremental Sort paths at additional places

  7. Fix representation of SORT_TYPE_STILL_IN_PROGRESS.

  8. Fix failures in incremental_sort due to number of workers

  9. Fix show_incremental_sort_info with force_parallel_mode

  10. Implement Incremental Sort

  11. Fix handling of "Subplans Removed" field in EXPLAIN output.

  12. Fix EXPLAIN (SETTINGS) to follow policy about when to print empty fields.

  13. Ensure plpgsql result tuples have the right composite type marking.

  14. Propagate sort instrumentation from workers back to leader.

  15. Make new regression test case parallel-safe, and improve its output.

  16. Push limit through subqueries to underlying sort, where possible.

  17. Fix inappropriate printing of never-measured times in EXPLAIN.

  18. Fix some infelicities in EXPLAIN output for parallel query plans.