Originally posted by: paul.brett

I've had some inquiries about why XML encoded strings get 'messed-up' when the data in question is parsed into an anyElem object.

Please review the enclosed example.

On input I have two types of data:

Data = Test & String

Stuff = any & elem

When reading this data into WTX, Data is read as XML, and the '&' is stripped back to just '&' and when mapped to a text object (see Output.txt) you get '&'.  Stuff is read as non-XML, and the '&' is not stripped back and when mapped to a text object (see Output.txt) you get '&'.

When mapping out to an XSD based card, the outputs are:

Data[1] = Test & String

Data[2] = <Stuff>any & elem</Stuff>

...and as you can see, all the XML elements have been encoded, but the '&' has been double-encoded to '&'.

This is not a bug, but a side-effect of using the anyElem feature of XML processing.

....

Can anybody suggest any elegant ways to avoid this double encoding?

My solution would be to pass the output to a RUN() map as a text blob, and do some SUBSTITUTE() functions to remove any double encoding, before writing it out.  Perhaps you know a better way?

Thank-you.

Paul.

Follow me on Twitter