Shape of passed values is (6, 4), indices imply (4, 4)

Question

I'm trying to make a Pandasandas DataFrame the data is zeros with shape (6,4) and the columns are

col=['Example', 'Example', 'Example', 'Example']

and the index is a list of lists:

ind=[[1,2,3,4],[1,2,3,4],[1,2,3,4],[1,2,3,4],[1,2,3,4],[1,2,3,4]]

and I'm doing it in this way:

pd.DataFrame(np.zeros((6,4)), columns = col, index=ind )

but it returns an error

Shape of passed values is (6, 4), indices imply (4, 1)

I tried to replace the list with a tuple and it worked! I'm wandering why this error is happening and how to solve it

score 0 · Accepted Answer · edited Mar 29 '23 at 12:16

0

You will need to pass an immutable object as index so that the index considers 6 indexes. You can do that by converting each of the sublists as tuples [tuple(i) for i in ind] -

ind=[[1,2,3,4],[1,2,3,4],[1,2,3,4],[1,2,3,4],[1,2,3,4],[1,2,3,4]]
col=['Example', 'Example', 'Example', 'Example']
pd.DataFrame(np.zeros((6,4)), columns = col, index=[tuple(i) for i in ind])

              Example  Example  Example  Example
(1, 2, 3, 4)      0.0      0.0      0.0      0.0
(1, 2, 3, 4)      0.0      0.0      0.0      0.0
(1, 2, 3, 4)      0.0      0.0      0.0      0.0
(1, 2, 3, 4)      0.0      0.0      0.0      0.0
(1, 2, 3, 4)      0.0      0.0      0.0      0.0
(1, 2, 3, 4)      0.0      0.0      0.0      0.0

You could create a multi-index out of this so that the index has 4 levels.

ind=[[1,2,3,4],[1,2,3,4],[1,2,3,4],[1,2,3,4],[1,2,3,4],[1,2,3,4]]
col=['Example', 'Example', 'Example', 'Example']
print(pd.DataFrame(np.zeros((6,4)), columns = col, index=pd.MultiIndex.from_tuples(ind)))

         Example  Example  Example  Example
1 2 3 4      0.0      0.0      0.0      0.0
      4      0.0      0.0      0.0      0.0
      4      0.0      0.0      0.0      0.0
      4      0.0      0.0      0.0      0.0
      4      0.0      0.0      0.0      0.0
      4      0.0      0.0      0.0      0.0

#Here first level is all 1 and second level is all 2 (and so on...)

edited Mar 29 '23 at 12:16

Marceli Wac

375
3
13

answered Feb 04 '21 at 13:09

Akshay Sehgal

18,741
3
21
51

is there any way to make the index as a list not a tuple – Karam Feb 04 '21 at 13:11
Nope , you can use multi-index however – Akshay Sehgal Feb 04 '21 at 13:12
The index has to be an immutable object and list is a mutable object. – Akshay Sehgal Feb 04 '21 at 13:14
Glad to help, do mark the answer if it helped you as it encourages me to assist you in the future as well :) – Akshay Sehgal Feb 04 '21 at 13:17
I will for sure – Karam Feb 04 '21 at 13:19
For more reading check this SO post as well. https://stackoverflow.com/questions/46193728/regarding-the-immutability-of-pandas-dataframe-indexes – Akshay Sehgal Feb 04 '21 at 13:19

Shape of passed values is (6, 4), indices imply (4, 4)

1 Answers1