[FIXED] zip two columns of a dataframe without changing the datatype

Issue

The problem is simple. Here we have a dataframe with a specified datatype for columns:

df = pd.DataFrame({'A':[1,2], 'B':[3,4]})
df.A = df.A.astype('int16')

#df
    A   B
0   1   3
1   2   4

#df.dtypes
A    int16
B    int64
dtype: object

Now I zip two columns A and B into a tuple:

df['C'] = list(zip(df.A, df.B))

    A   B   C
0   1   3   (1, 3)
1   2   4   (2, 4)

However, now the data type of values in column C are changed.

type(df.C[0][0])
#int
type(df.A[0])
#numpy.int16

How can I zip two columns and keep the datatype of each value inside the tuples, so that type(df.C[0][0]) would be int16 (same as type(df.A[0]))?

Solution

I think some type casting is happening when you refer as df.A, etc. See https://numpy.org/doc/stable/reference/generated/numpy.ndarray.tolist.html

Return a copy of the array data as a (nested) Python list. Data items are converted to the nearest compatible builtin Python type, via the item function.

But this worked

>>> import pandas as pd
>>> df = pd.DataFrame({'A':[1,2], 'B':[3,4]})
>>> df.A = df.A.astype('int16')
>>> df['C'] = list(zip(df.A.values, df.B.values))
>>> df
   A  B       C
0  1  3  (1, 3)
1  2  4  (2, 4)
>>> type(df.C[0][0])
<class 'numpy.int16'>
>>> type(df.C[0][1])
<class 'numpy.int64'>

Answered By - crayxt