[DFDL-WG] newbie question

Fri Mar 28 10:26:00 EDT 2014

Bing

Please be aware that DFDL is not designed to be a full blown data 
validator or rules engine. Rather, it is a data parser that can perform 
*some* validation, including XSDL validation (using XSDL facets) and some 
simple cross-field checking (using dfdl:assert). Note that the dfdl:assert 
mechanism is primarily intended to help the parser make the correct 
decisions, and not for validating.  That's partly why the DFDL expression 
language only uses a subset of XPath 2.0. If you need to perform complex 
cross-field validation (like checking that a 'total' field in a trailer 
record is the sum of the 'amount' fields in all the data records) then you 
are recommended to do that post-parse on the DFDL infoset that the parser 
creates, using a technology of your choice.

Regards

Steve Hanson
Architect, IBM DFDL
Co-Chair, OGF DFDL Working Group
IBM SWG, Hursley, UK
smh at uk.ibm.com
tel:+44-1962-815848

From:   Tim Kimber/UK/IBM at IBMGB
To:     "mfcplus at yahoo.com" <mfcplus at yahoo.com>, 
Cc:     dfdl-wg at ogf.org
Date:   28/03/2014 13:46
Subject:        Re: [DFDL-WG] newbie question
Sent by:        dfdl-wg-bounces at ogf.org

I didn't see the requirement for a pattern facet in your problem 
description. Feel free to use a pattern facet if you require one. 

If you want to perform cross-field validation then you should probably use 
DFDL asserts ( not discriminators ) for that. The best place to put the 
asserts is on the parent element, addressing ./A and ./C. 

You could use the exceptionType attribute on the assert (if the DFDL 
processor supports it) to make it throw a recoverable error instead of a 
processing error. That would enable invalid data to be parsed, and you 
would still be notified of any ( and all ) violations of the validation 
rules. 

regards,

Tim Kimber, 
IBM Integration Bus Development (Industry Packs)
Hursley, UK
Internet:  kimbert at uk.ibm.com
Tel. 01962-816742 
Internal tel. 37246742

From:        "mfcplus at yahoo.com" <mfcplus at yahoo.com> 
To:        Tim Kimber/UK/IBM at IBMGB, 
Date:        28/03/2014 13:21 
Subject:        Re: [DFDL-WG] newbie question 

so if i am not to use a pattern, how do i validate against the 
alphanumeric part? also i have a conditional in the message type, should i 
use discriminator with xpath rxpression for that? 
elemA /elemB/elemC CRLF elemC/elemD/elemE CRLF elemF/elemG/elemH CRLF 
for the above strcuture, i have a requirement that elemC must not exist if 
elemA is not there, and must exist if elemA is there. sequenceC must not 
exist if sequenceA is not there, sequenceC must exist if sequenceA is 
there. just wondering how do i do that with dfdl discriminator. mant 
thanks again
Sent from Yahoo Mail on Android. 

From: Tim Kimber <KIMBERT at uk.ibm.com>; 
To: mfcplus at yahoo.com <mfcplus at yahoo.com>; 
Cc: <dfdl-wg at ogf.org>; 
Subject: Re: [DFDL-WG] newbie question 
Sent: Fri, Mar 28, 2014 11:34:12 AM 
The key point here is that the data format is delimited, and there is a 
requirement to validate the parsed info set against some rules ( length 
constraints, in this case). 

I would 
- set terminator="/" 
- set lengthKind='delimited' 
- use xs:minLength and xs:maxLength facets to check the length ( or use a 
pattern facet if you must, but min/maxLength facets will use less CPU ) 
- switch on validation in the DFDL parser so that the length of the 
delimited value gets checked against the facets in the schema. 

Other possibilities exist ( e.g. DFDL assert to check the length ) but 
this is the most efficient. 

regards,

Tim Kimber, 
IBM Integration Bus Development (Industry Packs)
Hursley, UK
Internet:  kimbert at uk.ibm.com
Tel. 01962-816742 
Internal tel. 37246742

From:        "mfcplus at yahoo.com" <mfcplus at yahoo.com> 
To:        Tim Kimber/UK/IBM at IBMGB, 
Date:        28/03/2014 11:25 
Subject:        Re: [DFDL-WG] newbie question 

a field that is 3-5 chars long alphanumeric, basically a \w{3,5}, it 
should give out parsing error if it doesnt fit by the pattern, no 
initiator terminator is / 
Sent from Yahoo Mail on Android

From: Tim Kimber <KIMBERT at uk.ibm.com>; 
To: <dfdl-wg at ogf.org>; 
Subject: Re: [DFDL-WG] newbie question 
Sent: Fri, Mar 28, 2014 10:19:47 AM 

It depends on the data format. Normally a variable length field is 
terminated by a delimiter, so a data pattern/regex is not required. 
Can you describe the format in a little more detail? 

regards,

Tim Kimber, 
IBM Integration Bus Development (Industry Packs)
Hursley, UK
Internet:  kimbert at uk.ibm.com
Tel. 01962-816742 
Internal tel. 37246742

From:        Bing Lu <mfcplus at yahoo.com> 
To:        "dfdl-wg at ogf.org" <dfdl-wg at ogf.org>, 
Date:        28/03/2014 03:35 
Subject:        [DFDL-WG] newbie question 
Sent by:        dfdl-wg-bounces at ogf.org 

what's the best dfdl approache to describe  a variable length alpha 
numeric field? should i use simpletype restriction pattern, or use 
dfdl:assert testPattern? thanks in advance--
dfdl-wg mailing list
dfdl-wg at ogf.org
https://www.ogf.org/mailman/listinfo/dfdl-wg 

Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU

Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
--
  dfdl-wg mailing list
  dfdl-wg at ogf.org
  https://www.ogf.org/mailman/listinfo/dfdl-wg

Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ogf.org/pipermail/dfdl-wg/attachments/20140328/997aab28/attachment.html>