Matching value in a list to index where a condition is true

Asked May 14 '18 at 17:05

Active May 14 '18 at 17:11

Viewed 37 times

I have a set of columns that are either 1 or 0 depending on if a condition is true. Each row will have only a single column that is one.

I'm trying to add a new column to the df that finds the column number where the value equals 1 (e.g., column 0, 1, 2, etc.) and return the corresponding indexed value in a list that has the same number of column elements as the dataframe.

For example, if the value is 1 in the second column for row 1, I would want the newly added column to read b using the list below as a reference.

prim_cols = ['a','b','c','d','e','f']

What is the best python functions to use for this? I've tried np.where, list comprehension, and a few others, but I'm stuck.

edited May 14 '18 at 17:11

asked May 14 '18 at 17:05

J.B

2

I suspect you are trying to reverse an 1-hot encoding. This answer here might help you https://stackoverflow.com/questions/49372640/python-pandas-how-to-reverse-one-hot-encoding-back-to-categorical – kosnik May 14 '18 at 17:13
@kosnik: Which answer? :d – CristiFati May 14 '18 at 17:14
@kosnik Yes this is perfect, thank you! – J.B May 14 '18 at 17:28

Matching value in a list to index where a condition is true

0 Answers0