(Now that I'm on a real keyboard, more complete responses are a bit easier.)
Regarding the issue of missing columns from keep_col, you could create sets of what you have and what you want, and toss the rest: toss_these = list(set(df.columns) - set(keep_col)) del df[toss_these] Or something to that effect. Skip -- https://mail.python.org/mailman/listinfo/python-list
