Hi,

I need to sort XML documents in Microsoft Biztalk. I have implemented this using the XSLT sort.

Unfortunately I have found out that the sort is supposed to be based on the EBCDIC character set. The main difference between EBCDIC and, say, Unicode or ASCII, is that in EBCDIC letters come before numbers.

So I want:
ES00000122D7
USG07402AE39
US05967DAA90
US7615651004

not:
ES00000122D7
US05967DAA90
US7615651004
USG07402AE39

Any idea on how to do this. I have looked at the lang attribute but can't find any good resource on how to use it properly and if it will work for me in any case.

Thanks,

Michiel

I think Microsoft Biztalk is still using XSLT 1.0 so there's likely to be very little support for alternative collations, unless Microsoft have done something specific to their product. If you're only dealing with letters and digits, then a simple solution is to set up a translation table something like this:

You can obviously extend this to cover as many characters as you want. The first list should contain the characters in Unicode order, the second list in the order that you want them to collate (I think I've got that right...)

__________________
Michael Kay
http://www.saxonica.com/
Author, XSLT 2.0 and XPath 2.0 Programmer\'s Reference

Thanks a lot, it works like a charm.
I found out I actually need a sort where all small letters come before the capitals (abcdABCD where XSLT seems to do aAbBcCdD) but your method works for that as well.