[DDI-users] XSLT: How to iterate over contained var's or varGrp's

[Mark R. Diggory]

IF you can guarantee that the DDI's have a DTD or schema declaration in 
them, then I would recommend initially sticking with the ID/IDREF for 
XSLT 1.0. I've seen considerable discussion about having the keys 
function work to tokenize the list of IDREFs as a processing option for 
XSLT 2.0. But I think its far off.

Unfortunately ID/IDREF's become useless when the validation grammar is 
absent (ie when using a non-validating parser). Its not a bad idea have 
a robust solution that can work when the DTD/schema grammar/validation 
is absent from the parsed xml content.

As for using exslt functions, they can be very cross platform when it 
comes to xslt engines in java or any other language. They are a good 
choice in this regard. In many cases theres pure xslt versions of the 
functions like tokenize. For instance:

http://www.exslt.org/str/functions/tokenize/str.tokenize.template.xsl

<rant>

There's nothing worse than having a "Content Provider" generate DDI that 
doesn't include DTD/schema declarations. How do they guarantee this is 
actually a true DDI without them? The grammar and the associated 
namespace are truly the only thing that can guarantee that the document 
being parsed/transformed is truly a DDI document. At worst, I've seen 
developers write DDI content using standard text processing in perl 
without any use of an XML parser or grammar.

In most cases I fear, developers aren't even testing for such validation 
conditions when producing DDI content. This is as bad (if not worse) as 
the historical problems of poorly documented SPSS control cards and 
datasets with bad/absent metadata etc. If someones going to be producing 
DDI XML content, do the community a favor (and save your reputation in 
the process) include a DTD or Schema declaration. Your users will thank you.

</rant>

Andrew, better to make your code more robust to deal with such 
situations. Ultimately, the communities going to demand this be the case.

-Mark

Joachim Wackerow wrote:
> Andrew
> 
> With a small time lag I read by chance your question.
> 
> If an attribute is defined as IDREFS in the DTD like the attribute var 
> of the element varGrp and you are using the DDI DTD in the DDI document, 
> you could use the XPath function id() (see 
> http://www.w3c.org/TR/xpath#section-Node-Set-Functions).
> 
> It will give you a node-set, over which you could iterate, like:
> 
> <xsl:for-each select="id(/codeBook/dataDscr/varGrp/@var)">
>   <xsl:value-of select="labl"/>
> </xsl:for-each>
> 
> Achim
> 
> Andrew L Arnott wrote:
> 
>> Quick question on XSLT:
>>
>> If I have a varGrp with 5 var's referenced in it,
>> <varGrp ID="VG1" var="V1 V2 V3 V4 V5"/>
>>
>> And of course the necessary variables also in the XML, then how do I 
>> iterate
>> over those variables in XSLT?  I have used EXSLT to tokenize the 
>> string and
>> then do a search like this:
>> <xsl:variable name="IDlist"
>> select="ExsltStrings:tokenize(/codeBook/dataDscr/varGrp[@ID='VG1']"/>
>> <xsl:variable name="varList" 
>> select="/codeBook/dataDscr/var[@ID=$IDlist]"/>
>> <xsl:for-each select="varList">
>>     ...
>> </xsl:for-each>
>>
>> But I want to get away from using extension methods.  It seems like
>> something this common would be provided for in XSLT, maybe using 
>> xsl:key or
>> something like that.  But I just don't know it. Can someone please 
>> advise?
>>
>> Thanks,
>> Andrew
>>
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> DDI-users mailing list
>> DDI-users at icpsr.umich.edu
>> http://www.icpsr.umich.edu/mailman/listinfo/ddi-users
> 
> 
> _______________________________________________
> DDI-users mailing list
> DDI-users at icpsr.umich.edu
> http://www.icpsr.umich.edu/mailman/listinfo/ddi-users

-- 
Mark Diggory
Software Developer
Harvard MIT Data Center
http://www.hmdc.harvard.edu