DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta;
	h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references;
	b=WK6K5k3Vrrz+qch6jOHMQZB0BSu2Ff2NVFOKzfiKZB53PUEAjR2Mm3lma+coEmg6DbXJnR7k1Aq9lWfGbP4eunkRNjtclsEjFcEt0NfIPLE10LwCbpTZRKN+GjK2W+vGQm+36ES6e0gDc8GRLw0kdEeUsuKGyIncOBgigY5o2CU=
Message-ID: <758d5e7f0710030636m63bec244paa6cbaf04e66b3d2@mail.gmail.com>
Date: Wed, 3 Oct 2007 15:36:41 +0200
From: "Dawid Kuroczko" <qnex42@gmail.com>
To: weigelt@metux.de
Subject: Re: Finding broken regex'es
Cc: pgsql-sql <pgsql-sql@postgresql.org>
In-Reply-To: <20071002202313.GB18506@nibiru.local>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
References: <20071002202313.GB18506@nibiru.local>

On 10/2/07, Enrico Weigelt <weigelt@metux.de> wrote:
>
> Hi folks,
>
>
> I'm looking for some way to find broken regex'es in some column
> to kick them off. For now I'm regularily fetching all regexes
> from an PHP script, try an preg_match() and so find the broken
> ones to later remove them.
>
> Is there any way to do this directly within the db ?

Of course.  Exceptions is what You need!

CREATE FUNCTION regex_is_broken(r text) RETURNS boolean AS $$
    BEGIN
        PERFORM '' ~ r;
        RETURN 'f';
    EXCEPTION
        WHEN INVALID_REGULAR_EXPRESSION THEN
            RETURN 't';
    END;
$$ LANGUAGE PLpgSQL STRICT IMMUTABLE;

...and then you could do something like:

DELETE FROM table WHERE regex_is_broken(rx_col);


You don't need PLpgSQL to prevent such invalid regexes in the
first place.  You could use CHECK constraint for it:

CREATE TABLE rx_check (
    rx text CHECK ('' ~ rx IN ('t','f'))
);

postgres=> INSERT INTO rx_check (rx) VALUES ('.*');
INSERT 0 1
Time: 13.660 ms
postgres=> INSERT INTO rx_check (rx) VALUES ('234234');
INSERT 0 1
Time: 2.282 ms
postgres=> INSERT INTO rx_check (rx) VALUES ('par).*');
ERROR:  invalid regular expression: parentheses () not balanced

   Regards,
     Dawid