I'm trying to sum two columns: One is format float64 which are seconds and the other is format datetime64[ns]. They look like:
time | float
---------------------------------
2020-07-28 14:53:10 | 5.0
2021-08-30 15:28:52 | 68.0
2019-04-25 14:55:36 | 49.0
2020-03-24 09:19:45 | 8.0
2021-02-10 07:44:26 | 6.0
I performed the sum as follows:
df['time'] + df['float']
I then received the following message:
ufunc 'add' cannot use operands with types dtype('<M8[ns]') and dtype('float64')
I tried to convert the float column to timedate and then perform the sum and I received the following message:
unsupported operand type(s) for +: 'Timestamp' and 'Timestamp'`
Convert the "float" column to pd.Timedelta
using pd.to_timedelta
and then sum
:
df["time"] = pd.to_datetime(df["time"])
df["float"] = pd.to_timedelta(df["float"], unit="D") #change units if needed
>>> df.sum(axis=1)
0 2020-08-02 14:53:10
1 2021-11-06 15:28:52
2 2019-06-13 14:55:36
3 2020-04-01 09:19:45
4 2021-02-16 07:44:26
dtype: datetime64[ns]