I know this is an old thread (so forgive me for resurrecting it). But you could use a join that joins using the INPUT__FILE__NAME (meta date column) of the blob with some regexp cast on it (if neccesary) to make the name munchable.   So let's consider a table where you load the CSV data: CSV would read something like this: col1, col2, col3, col4, blob_loc some, random, data, in here. blob_loc some2, random2, data2, in here2, blob_loc   BLOB table would be <BLOB_content>, regexp_replace("", "", INPUT__FILE__NAME) as blob_loc   then just join on blob_loc   (typed this up in one minute, so forgive the formatting issues). ... View more