Numpy dstack Pandas dataframe columns

Asked May 01 '19 at 14:51

Active May 01 '19 at 14:56

Viewed 131 times

I want to take 2 pandas dataframe columns and combine them into 1.

Input: 3 Column df (columns: a , b , num)

Desired Output: 2 Column df (columns: (a,b), num)

Example df:

a b c 1 2 3 2 2 8 2 1 4

Example output:

a_b c (1,2) 3 (2,2) 8 (2,1) 4

df['a_b'] = np.dstack((df['a'], df['b']))

returns an error:

"Length of values does not match length of index"

Which I have tried to fix but do not fully understand since a, b, and num are all the same length, and the dstack version of a/b should also be that length.

What is going wrong with my approach and what would be the most understandable way to combine these two columns?

edited May 01 '19 at 14:56

asked May 01 '19 at 14:51

Joe S

Could you add a small example dataframe so we can see what your data looks like – Erfan May 01 '19 at 14:52
@Erfan done, good suggestion! – Joe S May 01 '19 at 14:57
3

`df['a_b'] = list(zip(df.a, df.b))` – user3483203 May 01 '19 at 15:10
This worked, I will google it to find out how, if you put it in an answer I will accept it, but I see that it is also documented here: https://stackoverflow.com/questions/45285244/convert-a-dataframe-to-list-of-tuples, should I delete this question? – Joe S May 01 '19 at 15:13
No it's a fine question @JoeS – Erfan May 01 '19 at 15:14

Numpy dstack Pandas dataframe columns

0 Answers0