Re: regexp_matches problem

Who many thanks Laurenz.
You gave me the right direction.
The following is what I made up and it does exactly what I want it to do:
    --collect the specific string data for each result row
SELECT (
    SELECT match[1] FROM
regexp_matches(secondresult.imgstring,'title="([^"]+)"','ig') AS match
    ) AS titletekst
    ,(
    SELECT match[1] FROM
regexp_matches(secondresult.imgstring,'alt="([^"]+)"','ig') AS match
    ) AS alttekst
    ,(
    SELECT match[1] FROM
regexp_matches(secondresult.imgstring,'src="([^"]+)"','ig') AS match
    ) AS imgsrc
FROM (
--collect the image string of every found image
    SELECT firstresult[1] AS imgstring
    FROM regexp_matches('<img src="wwww" title="titletext "
class="class12">'
        ,'<img[[:space:]]+([^>]*)>'
        ,'ig'
        ) AS firstresult
) as secondresult


This is it. Very simple and very efficient.

It took me ages to find out. Hopefully it helps other people.

Willem

-----Oorspronkelijk bericht-----
Van: pgsql-general-owner@postgresql.org
[mailto:pgsql-general-owner@postgresql.org] Namens Albe Laurenz
Verzonden: woensdag 13 mei 2009 10:50
Aan: WP Perquin *EXTERN*; pgsql-general@postgresql.org
Onderwerp: Re: [GENERAL] regexp_matches problem

WP Perquin wrote:
> When I make the following simplified example:
>
> SELECT regexp_matches('<img src="wwww" title="dit is een
> title tekst" class="class12">'
>
> ,'((title\s*=\s*\"([^"]*)")+)|((src\s*=\s*\"([^"]*)")+)','ig')
>
> My result are 2 rows:
>
> "{NULL,NULL,NULL,"src=\"wwww\"","src=\"wwww\"",wwww}"
>
> "{"title=\"dit is een title tekst\"","title=\"dit is een
> title tekst\"","dit is een title tekst",NULL,NULL,NULL}"
>
> I would like to have 1 row which contains both the records.
> Does anyone know how I can solve this?

Do you really want all those NULLs?

Is that what you want:

SELECT match[1]
FROM regexp_matches('<img src="wwww" title="dit is een title tekst"
class="class12">',
                    '(title\s*=\s*\"[^"]*"|src\s*=\s*\"[^"]*")',
                    'ig') AS match;

This query returns two rows.

When you write that you want to have one row that contains both
records, do you mean:

a) one string that is the concatenation of both strings
or
b) one row that is a single array with two string elements

Whatever it is you want, you will probably need to write a
little aggregate function that does that for you, something like

CREATE FUNCTION text_cats(state text[], nextv text) RETURNS text[]
   IMMUTABLE STRICT LANGUAGE sql
   AS 'SELECT $1 || $2';

CREATE AGGREGATE text_cat(text) (
   SFUNC = text_cats,
   STYPE = text[],
   INITCOND = '{}' );

for variant b).

Then you can

SELECT text_cat(match[1])
FROM regexp_matches('<img src="wwww" title="dit is een title tekst"
class="class12">',
                    '(title\s*=\s*\"[^"]*"|src\s*=\s*\"[^"]*")',
                    'ig') AS match;

Yours,
Laurenz Albe

--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general