changing mapping

I read in my record. As we proceed down the pipeline, I want to put kind of record (with length 742) into one path, and a different kind of record (with length 242) into a different path. My problem is that I have (obviously) different fixed width column mappings for the two kinds of records. How can I change the column mapping mid-stream? Can I? Do I need to just write out to flat files and go from there?

Thanks!

Jim Work
Yes, you have two options. Move each "path" to new, separate flat files (or database tables, or even RAW files) and then use the new files (or whatever chosen) in a new set of data flows to process the records.

The other option is once the record layouts have been split on their own paths is to use a derived column transformation to substring the data into the correct columns.|||Thanks a bunch!