Re: [HACKERS] Transform for pl/perl

From: Oleg Bartunov <obartunov(at)gmail(dot)com>
To: Anthony Bykov <a(dot)bykov(at)postgrespro(dot)ru>
Cc: Pavel Stěhule <pavel(dot)stehule(at)gmail(dot)com>, Pgsql Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: [HACKERS] Transform for pl/perl
Date: 2017-11-15 09:24:31
Message-ID: CAF4Au4wUZDKUheAQ0_5DEEPwc1cwEqg1PgvS1M+xBrm+UvafPA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 14 Nov 2017 11:35, "Anthony Bykov" <a(dot)bykov(at)postgrespro(dot)ru> wrote:

On Fri, 10 Nov 2017 14:40:21 +0100
Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com> wrote:

> Hi
>
> 2017-10-24 14:27 GMT+02:00 Anthony Bykov <a(dot)bykov(at)postgrespro(dot)ru>:
>
> > There are some moments I should mention:
> > 1. {"1":1}::jsonb is transformed into HV {"1"=>"1"}, while
> > ["1","2"]::jsonb is transformed into AV ["1", "2"]
> >
> > 2. If there is a numeric value appear in jsonb, it will be
> > transformed to SVnv through string (Numeric->String->SV->SVnv). Not
> > the best solution, but as far as I understand this is usual
> > practise in postgresql to serialize Numerics and de-serialize them.
> >
> > 3. SVnv is transformed into jsonb through string
> > (SVnv->String->Numeric).
> >
> > An example may also be helpful to understand extension. So, as an
> > example, function "test" transforms incoming jsonb into perl,
> > transforms it back into jsonb and returns it.
> >
> > create extension jsonb_plperl cascade;
> >
> > create or replace function test(val jsonb)
> > returns jsonb
> > transform for type jsonb
> > language plperl
> > as $$
> > return $_[0];
> > $$;
> >
> > select test('{"1":1,"example": null}'::jsonb);
> >
> >
> I am looking to this patch:
>
> 1. the patch contains some artefacts - look the word "hstore"
>
> 2. I got lot of warnings
>
>
> make[1]: Vstupuje se do adresáře
> „/home/pavel/src/postgresql/contrib/jsonb_plperl“
> gcc -Wall -Wmissing-prototypes -Wpointer-arith
> -Wdeclaration-after-statement -Wendif-labels
> -Wmissing-format-attribute -Wformat-security -fno-strict-aliasing
> -fwrapv -fexcess-precision=standard -g -ggdb -Og -g3
> -fno-omit-frame-pointer -fPIC -I../../src/pl/plperl -I. -I.
> -I../../src/include -D_GNU_SOURCE -I/usr/include/libxml2
> -I/usr/lib64/perl5/CORE -c -o jsonb_plperl.o jsonb_plperl.c
> jsonb_plperl.c: In function ‘SV_FromJsonbValue’: jsonb_plperl.c:83:9:
> warning: ‘result’ may be used uninitialized in this function
> [-Wmaybe-uninitialized] return (result);
> ^
> jsonb_plperl.c: In function ‘SV_FromJsonb’:
> jsonb_plperl.c:95:10: warning: ‘object’ may be used uninitialized in
> this function [-Wmaybe-uninitialized]
> HV *object;
> ^~~~~~
> In file included from /usr/lib64/perl5/CORE/perl.h:5644:0,
> from ../../src/pl/plperl/plperl.h:52,
> from jsonb_plperl.c:17:
> /usr/lib64/perl5/CORE/embed.h:404:19: warning: ‘value’ may be used
> uninitialized in this function [-Wmaybe-uninitialized]
> #define newRV(a) Perl_newRV(aTHX_ a)
> ^~~~~~~~~~
> jsonb_plperl.c:101:10: note: ‘value’ was declared here
> SV *value;
> ^~~~~
> gcc -Wall -Wmissing-prototypes -Wpointer-arith
> -Wdeclaration-after-statement -Wendif-labels
> -Wmissing-format-attribute -Wformat-security -fno-strict-aliasing
> -fwrapv -fexcess-precision=standard -g -ggdb -Og -g3
> -fno-omit-frame-pointer -fPIC -shared -o jsonb_plperl.so
> jsonb_plperl.o -L../../src/port -L../../src/common -Wl,--as-needed
> -Wl,-rpath,'/usr/lib64/perl5/CORE',--enable-new-dtags -Wl,-z,relro
> -specs=/usr/lib/rpm/redhat/redhat-hardened-ld
> -fstack-protector-strong -L/usr/local/lib -L/usr/lib64/perl5/CORE
> -lperl -lpthread -lresolv -lnsl -ldl -lm -lcrypt -lutil -lc make[1]:
> Opouští se adresář „/home/pavel/src/postgresql/contrib/jsonb_plperl“
>
> [pavel(at)nemesis contrib]$ gcc --version
> gcc (GCC) 7.2.1 20170915 (Red Hat 7.2.1-2)
> Copyright (C) 2017 Free Software Foundation, Inc.
> This is free software; see the source for copying conditions. There
> is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
> PARTICULAR PURPOSE.
>
> 3. regress tests passed
>
> 4. There are not any documentation - probably it should be part of
> PLPerl
>
> 5. The regress tests doesn't coverage other datatypes than numbers. I
> miss boolean, binary, object, ... Maybe using data::dumper or some
> similar can be interesting
>
> Note - it is great extension, I am pleasured so transformations are
> used.
>
> Regards
>
> Pavel
> >
> > --
> > Sent via pgsql-hackers mailing list (pgsql-hackers(at)postgresql(dot)org)
> > To make changes to your subscription:
> > http://www.postgresql.org/mailpref/pgsql-hackers
> >

Hello,
Thank you for your review. I have fixed most of your comments, except
for the 5-th part, about data::dumper - I just couldn't understand
your point, but I've added more tests with more complex objects if this
helps.

Please, take a look at new patch. You can find it in attachments to
this message (it is called "0001-jsonb_plperl-extension-v2.patch")

I'm curious, how much benefit we could get from this ? There are several
publicly available json datasets, which can be used to measure performance
gaining. I have bookmarks and review datasets available from
http://www.sai.msu.su/~megera/postgres/files/, look at js.dump.gz and
jr.dump.gz

--
Anthony Bykov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Pavel Stehule 2017-11-15 09:30:42 Re: [HACKERS] Transform for pl/perl
Previous Message Ildus Kurbangaliev 2017-11-15 09:09:28 Re: [HACKERS] Custom compression methods