convert pandas pd to numpy array and back

Question

What I want is to convert a numpy array to pandas dataframe.

df.head()
            A     B     C     D
      0    34    howdy  cow   meting
      1    23    cow    me    howdy

After tokenizing this df

df.head()
          A      B     C      D
     0    34     1     2      3
     1    23     2     4      1

converted df to numpy array for analysis with KMeans numpy array

   array [[34 ,1, 2, 3],
          [23 ,2, 4, 1]]

Question how can i convert this back to the first df i.e comparing the index of the array to the index of pandas and getting the row values

Answer 1

I think you can use values for convert to numpy array and then DataFrame constructor:

arr = df.values
print (arr)
[[34  1  2  3]
 [23  2  4  1]]

print (pd.DataFrame(arr))
    0  1  2  3
0  34  1  2  3
1  23  2  4  1
print (pd.DataFrame(arr, index=df.index, columns=df.columns))
    A  B  C  D
0  34  1  2  3
1  23  2  4  1