MIME-Version: 1.0
Date: Mon, 14 Aug 2023 10:04:03 -0400
From: Chapman Flack <chap@anastigmatix.net>
To: Andy Fan <zhihui.fan1213@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Pavel Stehule <pavel.stehule@gmail.com>,
 jian he <jian.universality@gmail.com>, pgsql-hackers
 <pgsql-hackers@postgresql.org>
Subject: Re: Extract numeric filed in JSONB more effectively
In-Reply-To: <CAKU4AWp+KLes8g=BWLqZfDmW9+=ZY0UC4G0i3qVcYEviK_dDTA@mail.gmail.com>
References: <CAKU4AWoqAVya6PBhn+BCbFaBMt3z-2=i5fKO3bW=6HPhbid2Dw@mail.gmail.com>
 <CACJufxHASOEpngQ8V2tbXgs4VZC3ETrVbS=uk0KC_B_J1j7ejQ@mail.gmail.com>
 <CAKU4AWrap1zpYqunJwWTN=CdP7E8e0U4mYmwn7hvTW3ERuENVg@mail.gmail.com>
 <CAFj8pRD-R-GsGCjeYApbhZoiW8TV6zACaYStMBMM0=--+WgN_A@mail.gmail.com>
 <CAKU4AWpDdFXAD+dMC1HeErXSKBUUBRGWkf=dAcX3wZgBNsWM=g@mail.gmail.com>
 <CAFj8pRAO3oEiBaJJ9=HZp6CoP2ffbwSgrKkLKjPfYZwx9wOOuQ@mail.gmail.com>
 <CAKU4AWoCHpKAVuQeOrk44cVPy_dVxn1aHrMUvHy5Ag-daFCSsQ@mail.gmail.com>
 <CAFj8pRD4cdUmK0RG4oN5B2KRSeDhwfMYaL=XpfEu4iaLeZ_Kow@mail.gmail.com>
 <CAKU4AWp8ab61e96v57OaB-Gm1bMfBNVLVy+s17U6_Ne3veB84g@mail.gmail.com>
 <3507485.1691090027@sss.pgh.pa.us>
 <CAKU4AWp+KLes8g=BWLqZfDmW9+=ZY0UC4G0i3qVcYEviK_dDTA@mail.gmail.com>
User-Agent: Roundcube Webmail/1.6.0
Message-ID: <900892914fdc3f477b101d699efb40e0@anastigmatix.net>
Content-Type: text/plain; charset=US-ASCII;
 format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: <https://www.postgresql.org/message-id/900892914fdc3f477b101d699efb40e0%40anastigmatix.net>
Precedence: bulk

On 2023-08-14 03:06, Andy Fan wrote:
>> We'd still have functions like jsonb_field_as_numeric() under the
>> hood, but there's not an expectation that users call them explicitly.
> 
> To avoid the lots of functions like jsonb_field_as_int2/int4, I defined
> Datum jsonb_object_field_type(.., Oid target_oid) at last,  so the
> function must return "internal" or "anyelement".
> ...
> I'm not sure how to fix that or deserves
> a fix? Or shall I define jsonb_object_field_int2/int8 to avoid this?

As far as I'm concerned, if the intent is for this to be a function
that is swapped in by SupportRequestSimplify and not necessarily to
be called by users directly, I don't mind if users can't call it
directly. As long as there is a nice familiar jsonb function the user
can call in a nice familiar way and knows it will be handled
efficiently behind the curtain, that seems to be good enough for
the user--better, even, than having a new oddball function to
remember.

However, I believe the rule is that a function declared to return
internal must also declare at least one parameter as internal.
That way, a user won't be shown errors about displaying its
returned value, because the user won't be able to call it
in the first place, having no values of type 'internal' lying
around to pass to it. It could simply have that trailing oid
parameter declared as internal, and there you have a strictly
internal-use function.

Providing a function with return type declared internal but
with no parameter of that type is not good, because then a
user could, in principle, call it and obtain a value of
'internal' type, and so get around the typing rules that
prevent calling other internal functions.

Regards,
-Chap