Re: Support tid range scan in parallel?

From: Cary Huang <cary(dot)huang(at)highgo(dot)ca>
To: "David Rowley" <dgrowleyml(at)gmail(dot)com>
Cc: "Pgsql Hackers" <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Support tid range scan in parallel?
Date: 2024-05-14 21:33:43
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

> You should add a test that creates a table with a very low fillfactor,
> low enough so only 1 tuple can fit on each page and insert a few dozen
> tuples. The test would do SELECT COUNT(*) to ensure you find the
> correct subset of tuples. You'd maybe want MIN(ctid) and MAX(ctid) in
> there too for extra coverage to ensure that the correct tuples are
> being counted. Just make sure and EXPLAIN (COSTS OFF) the query first
> in the test to ensure that it's always testing the plan you're
> expecting to test.

thank you for the test suggestion. I moved the regress tests from select_parallel
to tidrangescan instead. I follow the existing test table creation in tidrangescan
with the lowest fillfactor of 10, I am able to get consistent 5 tuples per page
instead of 1. It should be okay as long as it is consistently 5 tuples per page so
the tuple count results from parallel tests would be in multiples of 5.

The attached v4 patch includes the improved regression tests.

Thank you very much!

Cary Huang
HighGo Software Inc. (Canada)

Attachment Content-Type Size
v4-0001-add-parallel-tid-rangescan.patch application/octet-stream 21.3 KB

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Thomas Munro 2024-05-14 22:26:43 Re: Why is citext/regress failing on hamerkop?
Previous Message Jacob Burroughs 2024-05-14 21:21:40 Fwd: libpq compression (part 3)