[FIXED] pandas.to_datetime() returning different dtypes

Issue

Why does the code below gives rise to the following error UFuncTypeError: ufunc 'subtract' cannot use operands with types dtype('<M8[ns]') and dtype('float64')?

df1 = pd.DataFrame(['2022-01-01'], columns=['start'])
df1['start'] = pd.to_datetime(df1['start'])

df2 = pd.DataFrame(['2023-01-01'], columns=['end'])
df2['end'] = pd.to_datetime(df2['end'])

df2 - df1

Even though I can see that both arguments are showing datetime64[ns]:

print(df1.loc[0])
start   2022-01-01
Name: 0, dtype: datetime64[ns]

print(df2.loc[0])
end   2023-01-01
Name: 0, dtype: datetime64[ns]

Solution

You subtract first dataframe from second. If you want to get difference between date in dataframes, you need:

df2['end'] - df1['start']

Answered By - Kirill Kondratenko