Re: Running XSLT transformation

For example, it had silly attribute names like "textbox49" for the account title.

In the words of my Minnesota and Norwegian heritage: Oofta!

Can you expand on that?

Sure. Take the following XML:

<Order>
<Header orderDueDate="20080901" >
<Relationship id="134" />
</Header>
<Item>
...
</item>
</Order>

You might be tempted to stick the Relationship attribute id in with
the <Header> stuff during the XSL translation, but if the XSD says it
can repeat then you should be warned that it could (or most likely)
will break if they ever sent multiple <Relationship /> tags. Or it
might not break and you would just get the first or last id value
which would be even WORSE because you wouldn't know it was broken and
sometime down the road your partner would say "... yeah, didn't you
get the multiple Relationship id's"?

FWIW, the schema to which the existing document links is an invalid URL, so I'm not sure I can make it much worse :)

That URL is more than likely a qualifier to denote uniqueness vs.
being a document that is actually hosted at that exact URL location.
The only reason URL's (i.e. domain names) are employed in the naming
of a schema is to make it unique across all boundaries (i.e. across
all companies and countries). You should instead ask the other party
for a current XSD as it pertains to the document they are providing.

HTH,
Aaron Bartell
http://mowyourlawn.com



On Thu, Jul 10, 2008 at 12:36 PM, Adam Glauser <adamglauser@xxxxxxxxxxxx> wrote:

Aaron Bartell wrote:

I am not necessarily trying to steer you away from XSLT, but note that
you just introduced a potentially unnecessary failure point and
potentially CPU intensive processing that could be eliminated by just
just buckling down and doing it all in RPG.

I'm probably going to do it in RPG eventually, without the extra XSLT
step. The problem is that the vendor we're getting the data from seems
to have changed their web application, from which we are retrieving this
data. The XML is _really_ ugly, in that it's designed to be displayed
as a report as opposed to being designed for data transfer. For
example, it had silly attribute names like "textbox49" for the account
title. Not only that, but the account numbers are embedded in text
strings like "Account: ERB..... Montreal (QC)" instead of being an
account tag or attribute.

Another thing to watch out for is if you "flatten" the XML out, make
sure you are still adhering to the XSD (i.e. be mindful of repeating
sections).

Can you expand on that? I'm not using a validating parser. FWIW, the
schema to which the existing document links is an invalid URL, so I'm
not sure I can make it much worse :)

Just some thoughts that don't really answer your question (sorry) :-)

Appreciated nontheless.
--
This is the Websphere Development Studio Client for iSeries (WDSCI-L) mailing list
To post a message email: WDSCI-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/wdsci-l
or email: WDSCI-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/wdsci-l.