Reply to thread

Message: [QUOTE="Kam-Hung Soh, post: 3554078"] There doesn't seem to be a way to describe how specific columns should be processed in the csv module. You could define a conversion function that "guesses" the best conversion, for example: def str2num(datum): try: return int(datum) except: try: return float(datum) except: return datum for row in csv.reader(file(r'Transaction.csv')): [str2num(cell) for cell in row] ['Date', 'No.', 'Description', 'Debit', 'Credit'] ['3/17/2006', 5678, 'ELECTRONIC PAYMENT', '', 11.449999999999999] ['3/04/2007', 5678, 'THE HOME DEPOT 263 SomeCity FL', '', 25.399999999999999] [/QUOTE]