I am running into a similar problem and would like to get more info on this UDF.
In my case, we get data that at times included unprintable/invalid characters, mostly as a result of users pasting data from other documents (we get smart qoutes and things like that). Those characters cause problems when we query the data (primarily when a column is used in a Where clause).
I need an efficient way to replace these invalid characters.
I have problem in handling UNICODE colums with Fexp/MLoad. We use Named pipe access module to export and import data in parallel. Please share if any one has got the workaround for it. We use Teradata 12.