Add limit option to copy function

Started by 김대호over 6 years ago2 messageshackers
Beta feature

Hackorum builds and tests every patch posted to the lists, not only commitfest submissions. This is Hackorum's own CI rather than the PostgreSQL project's, and it is still under testing - please report anything that looks wrong.

won't retrysuccessCI history

You can run a PostgreSQL built from this patch straight from Docker, with no checkout and no build:

docker run --rm -p 5432:5432 ghcr.io/hackorum-dev/postgres-patch:t41816
psql -h localhost -U postgres

Built from patchset v1 (message #1), July 28, 2026 at 03:12 AM.

Every patchset is also pushed to a branch of our PostgreSQL fork, so you can check out the same tree CI built. Without a PostgreSQL checkout:

git clone --branch t41816_1 https://github.com/hackorum-dev/postgres.git

In a checkout you already have, add the fork once:

git remote add hackorum https://github.com/hackorum-dev/postgres.git

then, for this patchset and every later one:

git fetch hackorum t41816_1 && git checkout t41816_1

Patchset v1 (message #1) is on t41816_1

Jump to latest
#1김대호
daiho1.kim@samsung.com

Attachments:

t41816_1
copy-limit-feature.patchapplication/octet-streamDownload+32-0
#2Tom Lane
tgl@sss.pgh.pa.us
In reply to: 김대호 (#1)
Re: Add limit option to copy function

=?UTF-8?B?6rmA64yA7Zi4?= <daiho1.kim@samsung.com> writes:

I suggest adding a limit option to the copy function that limits count of input/output.
I think this will be useful for testing with sample data.

I'm quite skeptical of the value of this. On the output side, you
can already do it with

COPY (SELECT ... LIMIT n) TO wherever;

Moreover, that approach allows you to include an ORDER BY, which is
generally good practice in any query that includes LIMIT, in case
you'd like deterministic results.

On the input side, it's true that you'd have to resort to some
outside features (perhaps applying "head" to the input file, or
some such), or else copy the data into a temp table and post-process.
But that's true for most ways that you might want to adjust or
filter the input data; why should this one be different?

We don't consider that COPY is a general-purpose ETL engine, and
have resisted addition of features to it in the past because
they'd slow down the primary use-case. That objection applies
here too. Yeah, it's (probably) not a big slowdown ... but it's
hard to justify any cost at all for a feature that is outside
the design scope of COPY.

regards, tom lane