[BUG] DType of Elements in Lists changes

[mlahir1]

DType of Elements in Lists changes with df1['trans'] = df1['xyz']

Repro:

>>> import cudf
>>> df =cudf.DataFrame()
>>> df['unt_qty'] = [['a', 'b'], []]
>>> df['trans'] = [1, 2]
>>> df['xyz'] = [1121212,21231232]
>>> df1 = df[df.trans.isna()]
>>> df1['trans'] = df1['xyz']
>>> df1.unt_qty.dtype
ListDtype(object)

>>> df1 = df1.drop(['xyz'], axis=1)
>>> df1.unt_qty.dtype
ListDtype(int8)

if i comment out the line df1['trans'] = df1['xyz'] the dtype of list remains same.

>>> import cudf
>>> df =cudf.DataFrame()
>>> df['unt_qty'] = [['a', 'b'], []]
>>> df['trans'] = [1, 2]
>>> df['xyz'] = [1121212,21231232]
>>> df1 = df[df.trans.isna()]
>>> df1.unt_qty.dtype
ListDtype(object)

>>> df1 = df1.drop(['xyz'], axis=1)
>>> df1.unt_qty.dtype
ListDtype(object)

Impact: This causing issue in concats.