Is it possible to have a custom reader that creates multiple datasets in the pipeline versus have the data read as a MultiBlockDataSet then having to run the "ExtractBlock" filter? Thanks -- Mike Jackson imikejackson & gmail * com