pandas.read_excel（）与标题行列表

Question

I'm trying to parse an Excel file with a table which has two header rows looking like:

| Category A | Category B |
|  1   |  2  |  1   |  2  |

According to the manual I can pass a list of ints to specify rows to use for a multi-index, but when I do:

pd.read_excel(path, header = [1,2])

I get back an error saying "list indices must be integers, not list". Apparently I'm not allowed to put any list as header kwarg, as I get the same error when doing header = [1] (as opposed to header=1 , which works fine).

What am I doing wrong here?

Answer 1

It sounds like you might not have the Pandas 0.17.0 installed (the man page you linked to).

The read_csv function has been changed - Pandas 0.14.0 (for example) didn't allow a list of ints to be passed as a header argument, see here .

Answer 2

First, the rows you put into header keyword should be 0-indexed, so you probably want header=[0,1] ?

Second, there was a bug if the multi-index columns contained numeric values, which will be fixed in 0.17.1 (coming soon), or you can build from master - https://github.com/pydata/pandas/pull/11328