One method of dealing with troublesome textfiles is to load them as fixed record files without specifying field breaks. Each record will then contain a single field holding the entire row.
The next step is to use the SubField() function, using the | delimiter, in a resident load to split up the data in fields. Hopefully you'll be able to load most data this way. If you have isolated the problem you may be able to do some damage control here on the problem rows as well.
You will need to specify each field (field1, field2...) of course
You probably have a missing '|' field separator somewhere in your records.
I know it's very annoying to debug this kind of issues, but you can load whole line from the text file as one text field and calculate a number of '|' characters in each one of them.
The string(s) with anomaly (with the number of separators less than usual) should be pointed out easily.
Also if I load the offending line on it's own with just the header row it loads fine. It's as if the structure of the file itself is what's causing the problem
