Re: [v10] CREATE TEMP FUNCTION/CREATE FUNCTION PG_TEMP.X

From: Vincenzo Romano <vincenzo(dot)romano(at)notorand(dot)it>
To: Melvin Davidson <melvin6925(at)gmail(dot)com>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, PostgreSQL General <pgsql-general(at)postgresql(dot)org>
Subject: Re: [v10] CREATE TEMP FUNCTION/CREATE FUNCTION PG_TEMP.X
Date: 2017-12-22 07:33:40
Message-ID: CAHjZ2x5Zgf0eiX0M4bZY1XnL47_3MPJPzJXhjV07sAiuk-SHAQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

2017-12-22 0:50 GMT+01:00 Melvin Davidson <melvin6925(at)gmail(dot)com>:
>
>
>
> On Thu, Dec 21, 2017 at 12:07 PM, Vincenzo Romano <vincenzo(dot)romano(at)notorand(dot)it> wrote:
>>
>> 2017-12-21 17:56 GMT+01:00 Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>:
>> > Vincenzo Romano <vincenzo(dot)romano(at)notorand(dot)it> writes:
>> >> 2017-12-21 17:52 GMT+01:00 Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>:
>> >>> You have to schema-qualify the temp function name when calling it, too.
>> >
>> >> So search_path is not used with functions?
>> >
>> > pg_temp is explicitly ignored when searching for functions/operators.
>> > Otherwise, installing a trojan horse is just too easy.
>> >
>> > regards, tom lane
>>
>> I'm not sure whether this decision actually makes PG more scure.
>> But, anyway, thanks for the insight: I've just found the
>> documentations for this.
>>
>> --
>> Vincenzo Romano - NotOrAnd.IT
>> Information Technologies
>> --
>> NON QVIETIS MARIBVS NAVTA PERITVS
>>
>
> Aside from the simple explanations you have received, I question your justification for even having a temporary function.
> Functions are only entries in the system catalogs and as such, take up just a tiny amount of physical space. In addition,
> if you ever need it again, you will have to expend time recreating it. Why not just once and keep it?
>

Hi.
Thanks for your comment.

The reason for having temporary object, in my current design, is to
have something shadowing something else on a per session basis, thanks
to the search_path variable.
It's not simply a matter or storage room or access speed. Not at all to me.
If you use, for example:

SET search_path to pg_temp,"$user",public;

you can put general stuff in public, per-user data in "$user" and per
session data in pg_temp.
Then the "name resolution" will follow the above priority during lookup.
And, as I put more and more logics in the DB, having temporary
functions gives me a simple, clean and yet powerful design.

As soon as my applications connect, they run SELECT * FROM
f_application_init( 'MYAPPNAME' ).
That function (which is not temporary) will setup the DB-level, the
user-level and the session-level stuff.
Currently it eats about 500 msec to run and it's run only once per session.
So, the answer to your question is: "why not if it can be useful
"
--
Vincenzo Romano - NotOrAnd.IT
Information Technologies
--
NON QVIETIS MARIBVS NAVTA PERITVS

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Ken Tanzer 2017-12-22 07:56:23 Re: Intersection or zero-column queries
Previous Message Tom Lane 2017-12-22 02:59:46 Re: Intersection or zero-column queries