Re: [PATCH] expand the units that pg_size_pretty supports on output - Mailing list pgsql-hackers

From David Rowley
Subject Re: [PATCH] expand the units that pg_size_pretty supports on output
Date
Msg-id CAApHDvoWuEtD8ec3CkEZXQn6m2ZFBe5MU4gm226ksZyUW5LPKQ@mail.gmail.com
Whole thread Raw
In response to Re: [PATCH] expand the units that pg_size_pretty supports on output  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On Thu, 8 Jul 2021 at 07:31, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> David Christensen <david.christensen@crunchydata.com> writes:
> > Enclosed is the patch to change the return type to numeric, as well as one for expanding units to
> > add PB and EB.
>
> Can we really get away with changing the return type?  That would
> by no stretch of the imagination be free; one could expect breakage
> of a few user views, for example.

That's a good point. We should probably leave it alone then.  I had
had it in mind that it might be ok since we did this for extract() in
14. At least we have date_part() as a backup there. I'm fine to leave
the return value of pg_size_bytes as-is.

> Independently of that, I'm pretty much -1 on going further than PB.
> Even if the additional abbreviations you mention are actually recognized
> standards, I think not that many people are familiar with them, and such
> input is way more likely to be a typo than intended data.

I'm fine with that too.  In [1] I mentioned my concerns with adding
all the defined units up to Yottabyte. David reduced that down to just
exabytes, but I think if we're keeping pg_size_bytes returning bigint
then drawing the line at PB seems ok to me. Anything more than
pg_size_bytes('8 EB') would overflow.

David

[1] https://www.postgresql.org/message-id/CAApHDvp9ym+RSQNGoSRPjH+j6TJ1tFBhfT+JoLFf_RbZq1EszQ@mail.gmail.com



pgsql-hackers by date:

Previous
From: Filip Janus
Date:
Subject: SHA-1 FIPS - compliance
Next
From: Masahiko Sawada
Date:
Subject: Small documentation improvement for ALTER SUBSCRIPTION