Pandas overwrite column names by position

Question

I'm reading in a csv, and I'd like to overwrite the first two column names.

• I can't use df.rename(columns={0:'name 1', 1:'name 2'}) because the columns aren't called 0 and 1. They have names, I just want to throw them out.

• It seems like

  df.columns.values[0] = 'name 1'
  df.columns.values[1] = 'name 2'
  

  has serious issues because afterwards, df['name 1'] gives me a KeyError.

• What would be ideal would be pd.read_csv(file, names=['name 1', 'name 2', ...]. Curiously, this renames col 3 "Ellipsis" and doesn't have the desired effect.

Any ideas for how to do this sensibly in pandas?

Answer 1

Try this:

df.columns = ['name 1', 'name 2'] + df.columns[2:].tolist()