3D numpy array to dataframe since pandas panel has been deprecated

Question

x = np.random.rand(100, 4, 4)
x

   [[[0.71, 0.65, 0.78, 0.85],
    [0.09, 0.43, 0.41, 0.28],
    [0.25, 0.12, 0.8 , 0.44],
    [0.42, 0.06, 0.63, 0.26]],
   ...,

   [[0.46, 0.86, 0.13, 0.51],
    [0.37, 0.2 , 0.79, 0.13],
    [0.39, 0.3 , 0.34, 0.77],
    [0.48, 0.71, 0.38, 0.39]]])

I want to convert this into a pandas data frame such that

x[0][:,0] is column1, 
x[1][:,0] is column2 
and so on.

I searched for this. However all the answers I could find use PD.panel which has been deprecated.

I could think of for-loop, but maybe with larger dimension could be very slow. — Karina
– Karina, Commented Aug 19, 2021 at 6:21

Anurag Dabas · Accepted Answer · 2021-08-19 06:34:52Z

1

You can use hstack():

df=pd.DataFrame(np.hstack(x))

OR

use concatenate()

df=pd.DataFrame(np.concatenate(x,axis=1))

output of df:

      0          1            2      ...      397           398     399
0   0.095396    0.963393    0.218431 ...    0.369935    0.094349    0.536987
1   0.742115    0.641849    0.555946 ...    0.732853    0.946447    0.987616
2   0.825148    0.937934    0.013133 ...    0.595918    0.486560    0.816795
3   0.049911    0.214139    0.702705 ...    0.406167    0.872152    0.488658

Note: you can also use reshape() but it will messup the order:

df=pd.DataFrame(x.reshape(x.shape[-1],x.shape[0]*x.shape[-1]))

edited Aug 19, 2021 at 6:34

answered Aug 19, 2021 at 6:22

Anurag Dabas

24.3k9 gold badges25 silver badges41 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

3D numpy array to dataframe since pandas panel has been deprecated

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related