Horton

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

Alvaro Herrera <alvherre@2ndquadrant.com>

View thread View on postgr.es
From: Alvaro Herrera <alvherre@2ndquadrant.com>
To: James Coleman <jtc331@gmail.com>
Cc: Tomas Vondra <tomas.vondra@2ndquadrant.com>, 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>, Andreas Karlsson <andreas@proxel.se>
Date: 2020-03-24T03:44:38Z
Lists: pgsql-hackers 
On 2020-Mar-23, James Coleman wrote:

> 4. nodeIncrementalSort.c ExecReScanIncrementalSort: This whole chunk
> is suspect. I've mentioned previously I don't have a great mental
> model of how rescan works and its invariants (IIRC someone said it was
> about moving around a result set in a cursor). Regardless I'm pretty
> sure this code just doesn't work correctly.

I don't think that's the whole of it.  My own vague understanding of
ReScan is that it's there to support running a node again, possibly with
different parameters.  For example if you have a join of an indexscan
on the outer side and an incremental sort on the inner side, and the
values from the index are used as parameters to the incremental sort,
then the incremental sort is going to receive ReScan calls for each of
the values that the index returns.  Sometimes the index could give you
the same values as before (because there's a dupe in the index), so you
can just return the same values from the incremental sort; but other
times it's going to return different values so you need to reset the
incremental sort to "start from scratch" using the new values as
parameters.

Now, if you have a cursor reading from the incremental sort and fetch
all tuples, then rewind completely and fetch all again, then that's
going to be a rescan as well.

I agree with you that the code doesn't seem to implement that.


-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

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.