Handling double-byte data values in Mercator/DataStage TX

We are getting ANSI X12 EDI files that has double-byte(Japanese,
Chinese,... characters.) data in some of the data elements and we are
mapping it to xml fields. We are using DataStage TX 7.5 for data
mapping.

I would like to compare the double-byte data values in the EDI
elements and conditionally map in output xml fields.

Can anyone give idea how we can handle this type of requirement in
Mercator maps/typetrees or ...?

Any information on this is greatly appreciated.

Thanks,
Kiran