What is the best way in python to get the column number of the max values of each row in a pandas dataframe-matrix (without the index column). E.g. if I would have
| Date | Company 1 | Company 2 | Company 3 |
|---|---|---|---|
| 01.01.2020 | 23 | 21 | 14 |
| 02.01.2020 | 22 | 12 | 22 |
| 03.01.2020 | 11 | 11 | 12 |
| .... | ... | ... | ... |
| 02.01.2020 | 2 | 14 | 3 |
The output should be the vector:
[1, 1, 3, ..., 2]
CodePudding user response：
Use idxmax:
In [949]: cols = df.iloc[:, 1:].idxmax(1).tolist()
In [950]: cols
Out[950]: ['Company 1', 'Company 1', 'Company 3']
If you want column index:
In [951]: [df.columns.get_loc(i) for i in cols]
Out[951]: [1, 1, 3]