I spend a lot of time writing reports from flat datafiles. One thing about the csv.c extension is that it's FAST even when creating a table from the virtual table. My challenge is that I often create TSV in addition to CSV. I've done the batch sed/awk thing in the past but it seems to me that either forking the csv.c for a tsv.c or adding a `sep=""` param would make more sense. As I'm not qualified to write SQLite internals I was hoping for some guidance. Forking seems to add a lot of code where a param requires more test cases... and I may not be prepared for that either. All that said, some guidance and then we'll see what I can do. regards