Re: [HACKERS] xpath not a good replacement for xpath_string

On Tue, Jul 28, 2009 at 3:21 PM, <pgsql@xxxxxxxxxxxxxx> wrote:
> Sorry to bring this up, I know you've been fighting about XML for a while.
>
> Currently, I am using XML2 functionality and have tried to get the newer
> XPath function to work similarly, but can't quite seem to do it.
>
> I think the current xpath function is too limited. (The docs said to post
> problems to hackers if I have an issue.)
>
> For instance, we have a web application that uses java with an XML class
> serializer/deserializer Xstream. It creates XML that looks like this:
>
> <com.company.local.myclass>
>    <uuid>b5212259-a91f-4dca-a547-4fe89cf2f32c</uuid>
>    <email>joe@xxxxxxxxxxxxxx</email>
> </com.company.local.myclass>
>
> My current strategy is to use xml2 as:
>
> select xpath_string(datum, E'/com\.company\.local\.myclass/uuid) as uuid
> from table;
>
> Which produces a usable:
> b5212259-a91f-4dca-a547-4fe89cf2f32c
>
> I have been trying to use xpath
> select xpath(E'/com\.company\.local\.myclass/uuid', XMLPARSE(CONTENT
> datum)) as uuid from table;
>
> Which produces an unusable:
> {<uuid>b5212259-a91f-4dca-a547-4fe89cf2f32c</uuid>}
>

How about:

SELECT (xpath(E'/com\.company\.local\.myclass/uuid/text()',
XMLPARSE(CONTENT datum)))[1] as uuid from table;

Not as clean, but it produces the same result as xpath_string().
Combined with array_to_string() could can collapse the array instead
of just grabbing the first element (in cases other than uuid, of
course).

-- 
Mike Rylander
 | VP, Research and Design
 | Equinox Software, Inc. / The Evergreen Experts
 | phone:  1-877-OPEN-ILS (673-6457)
 | email:  miker@xxxxxxxxxxxxxx
 | web:  http://www.esilibrary.com

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