I've just done the same thing. My inputs are CSV files from all over the place, and the output is a ledger file. Driven by make, of course. :) Processing the C

I've just done the same thing. My inputs are CSV files from all over the place, and the output is a ledger file. Driven by make, of course. :) Processing the CSVs takes about one second per year of my data, so what I think I'll probably do is summarize it out to ledger format each year. I was using awk to do the CSV parsing, and it's blazing fast. But it gets a little bit unwieldy if you have to do multiple passes through a file or if a single CSV file has multiple sections. So I moved them to Python which is far more capable, but slower.