multicolumn data store filter

I created a multicolumn data store on 20 blob columns.  I set the DELIMITER attribute to NEWLINE, and set the FILTER attribute to Y for all 20 blobs.  This is because each of the blobs is read in from a file, which can be text, MS Word, MS excel, etc, and I do not necessarily know which of the 20 blobs has which kind of data.

My problem is that one of these files is plain text, including English and Arabic characters.  The resync of the text index fails, with

DRG-11207: user filter command exited with status 1

DRG_11222: Third-party filter does not supprt this known document format.