Re: Proposal: QUALIFY clause - Mailing list pgsql-hackers

From Thom Brown
Subject Re: Proposal: QUALIFY clause
Date
Msg-id CAA-aLv7D7Bes=4f4_=a7k3rq-oLdw2y1ama6voLbF5yUphnOWQ@mail.gmail.com
Whole thread Raw
In response to Re: Proposal: QUALIFY clause  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: Proposal: QUALIFY clause
List pgsql-hackers
On Mon, 21 Jul 2025, 18:31 Tom Lane, <tgl@sss.pgh.pa.us> wrote:
"Matheus Alcantara" <matheusssilv97@gmail.com> writes:
> You're right — semantically, using QUALIFY is equivalent to wrapping the
> query in a subquery and applying a WHERE clause to the result. The main
> motivation here is to provide a more ergonomic and readable syntax.

> While I understand the hesitation around introducing another keyword
> that effectively acts like WHERE at a different stage, I believe QUALIFY
> improves clarity in many use cases, by avoiding the boilerplate and
> visual noise of nested subqueries making it easier to write and reason
> about.

There are concrete reasons not to do this until/unless it becomes
standardized:

* If the syntax is like WHERE, there will be no way to do it without
making QUALIFY a fully-reserved word.  That will inevitably break
more than zero applications.  It's a lot easier to justify that
sort of breakage if we can say "QUALIFY is reserved according to
SQL:20xx, so don't blame us".

* I'm not exactly convinced that the committee would standardize
it just like this.  For one thing, QUALIFY is not even the right
part of speech: it's a verb, and thus more fit to be a primary
statement keyword.  What you need here is an adverb (I think ...
been a long time since high school English, but my dictionary
says WHERE is an adverb).  Maybe they'd be persuaded to do what
the existing implementations did, but I wouldn't be at all surprised
if they choose a different keyword.

I know we're not bikeshedding, but the word REFINE might be more appropriate.

Thom

pgsql-hackers by date:

Previous
From: "Vitale, Anthony, Sony Music"
Date:
Subject: Question on any plans to use the User Server/User Mapping to provide Logical Replication Subscriptions the user/password in an encrypted manner
Next
From: Andres Freund
Date:
Subject: Re: [PATCH] Let's get rid of the freelist and the buffer_strategy_lock