Re: [xsl] How to "flatten" a XML file with n-ary Cartesian product over n sequences X1, ..., Xn using XSL

[Martin Honnen <Martin.Honnen@xxxxxx>]

reynaldo.rizzo@xxxxxxxxxxxxxxxx wrote:

> The desired output is a delimited file shown as follows:
>
> 'output.csv'
> root,itemCode,itemName,aCode,aDescription,bCode,bDescription,cCode,cDescription

That line has nine items while the lines below have only eight items each.

> 01,name0,10,description0,100,description2,996,description4
> 01,name0,10,description0,100,description2,997,description5
> 01,name0,10,description0,200,description3,996,description4
> 01,name0,10,description0,200,description3,997,description5
> 02,name1,20,description6,null,null,998,description10
> 02,name1,20,description6,null,null,999,description11
>
> Output is the Cartesian product of each item with its own sub-sequences.
>
> I've had limited success with my own limited self-learned XSL skills 
> dealing on a per case basis, but I'm not satisfied with the efficiency 
> achieved. Because I have no "normalized" approach, complexity of my XSL 
> grows considerably, related to the number of sequences involved and I'm 
> getting stuck when it involves more than three.
>
> I was wondering if there is any "standard" method that solves this problem 
> for n-sequences (i.e. using a <xsl:for-each-group> that groups by item and 
> repeat the contents for each sublist?) so I could use recursion or 
> iterations according to n. Please note that sequences are optional; the 
> schema allows for empty or nonexistent sequences, where in such cases 
> 'null' must be output for the missing fields.

And what determines the complete number of sequences?

--

	Martin Honnen
	http://msmvps.com/blogs/martin_honnen/