[DFDL-WG] Proposed feature: lookup tables via simple type unions

[Steve Hanson]

Hi Mike

I have taken a look at your revised proposal at 
https://cwiki.apache.org/confluence/display/DAFFODIL/Proposal%3A+Features+to+Support+Table-Lookup
.

Can you augment the example with use of the simple type by an element?   I 
think this might highlight a problem. Today we allow the DFDL 
representation properties for a simple element to be split across the 
element declaration and the (hierarchy) of simple types.  Your proposal is 
allowing the rep type of an element to be other than the allowed set for 
the simple type of the element.  This prevents the combination of rep 
properties across the element and the simple type. In fact, it means that 
all type dependent rep properties would have to be carried by the rep 
type. That's not going to work because of scoping rules - you can't just 
switch off property applicability. 

Regards
 
Steve Hanson
IBM Hybrid Integration, Hursley, UK
Architect, IBM DFDL
Co-Chair, OGF DFDL Working Group
smh at uk.ibm.com
tel:+44-1962-815848
mob:+44-7717-378890
Note: I work Tuesday to Friday 



From:   Steve Hanson/UK/IBM
To:     Mike Beckerle <mbeckerle.dfdl at gmail.com>
Cc:     "dfdl-wg at ogf.org" <dfdl-wg at ogf.org>
Date:   04/04/2017 13:42
Subject:        Re: [DFDL-WG] Proposed feature: lookup tables via simple 
type unions


Although signed-up to Confluence I couldn't see a way to comment on the 
proposal there.

I'm not comfortable with the proposal.

1)  It relies upon validation being enabled, which is an optional feature. 
There is a long-held principle that switching on validation does not 
change the behaviour of a parse, but this will do exactly that. 

[unionMemberSchema] String. For simple element information items, this 
member contains an SCD reference to the member of the union that matched 
the value of the element. Empty if validation is not enabled. Empty if the 
element's type is not a union.
2) DFDL does not allow the concept of an annotation on a simple type that 
is a union member. We would be allowing that, but with a completely 
disjoint property set. 
Unions; the memberTypes must be derived from the same simple type. DFDL 
annotations are not permitted on union members

3) The proposal does not help the case where I am not using unions but 
still would like enums translated into meaningful strings, a far more 
common situation.

I think this processing is best left to either an independent post-parse 
step or a parser extension via a set of non-DFDL annotations. 

Regards
 
Steve Hanson
IBM Hybrid Integration, Hursley, UK
Architect, IBM DFDL
Co-Chair, OGF DFDL Working Group
smh at uk.ibm.com
tel:+44-1962-815848
mob:+44-7717-378890




From:   Mike Beckerle <mbeckerle.dfdl at gmail.com>
To:     "dfdl-wg at ogf.org" <dfdl-wg at ogf.org>
Date:   22/02/2017 17:38
Subject:        [DFDL-WG] Proposed feature: lookup tables via simple type 
unions
Sent by:        "dfdl-wg" <dfdl-wg-bounces at ogf.org>




A write up of the proposal, which we are prototyping in Daffodil, is here:

https://opensource.ncsa.illinois.edu/confluence/display/DFDL/Enumerations+and+Range+Tables+via+Simple+Type+Unions


This is needed by a number of data formats we are working with where there 
are large enumerations having as many as 2000 members. Often these 
enumerations are a mixture where single values correspond to some 
enumerated strings, and ranges of  values correspond to others.  

Using expressions to translate representation integers into strings is 
infeasible, as no constant-time case-statement-like construct is available 
in DFDL.

Mike Beckerle | OGF DFDL Workgroup Co-Chair | Tresys Technology | 
www.tresys.com
Please note: Contributions to the DFDL Workgroup's email discussions are 
subject to the OGF Intellectual Property Policy
--
  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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ogf.org/pipermail/dfdl-wg/attachments/20180912/b9540606/attachment-0001.html>