[FIXED] I want to create a new column by converting the timestamp: AttributeError: Can only use .str accessor with string values

Issue

df = {'ArrivalPeriod': [2018-10-30 06:49:57, 2019-11-30 07:42:53, 2018-11-28 02:21:57, 2018-10-30 06:49:57],
'time': [12, 8, 10, 10],
'weekday': [4, 6, 5, 1]}
df = pd.DataFrame(df)
df

df['Arrival_Time'] = pd.DatetimeIndex(df['ArrivalDate']).time
df['Arrival_Date'] = pd.DatetimeIndex(df['ArrivalDate']).Date

df["Arrival_Time"] = df["Arrival_Time"].str.split(':').apply(lambda x: int(x[0]) * 60 + int(x[1]))

My expected output is to convert the time seconds.

Solution

Try using .dt:

df = pd.DataFrame({'ArrivalPeriod': ["2018-10-30 06:49:57",  "2019-11-30 07:42:53",  "2018-11-28 02:21:57",  "2018-10-30 06:49:57"],
                   'time': [12, 8, 10, 10],
                   'weekday': [4, 6, 5, 1]})

df["ArrivalPeriod"] = pd.to_datetime(df["ArrivalPeriod"])
df["Arrival_Time"] = df["ArrivalPeriod"].dt.time
df["Arrival_Date"] = df["ArrivalPeriod"].dt.date

>>> df
        ArrivalPeriod  time  weekday Arrival_Time Arrival_Date
0 2018-10-30 06:49:57    12        4     06:49:57   2018-10-30
1 2019-11-30 07:42:53     8        6     07:42:53   2019-11-30
2 2018-11-28 02:21:57    10        5     02:21:57   2018-11-28
3 2018-10-30 06:49:57    10        1     06:49:57   2018-10-30

If you want to get the total seconds from the time, you can then do:

>>> pd.to_timedelta(df["Arrival_Time"].astype(str)).dt.total_seconds()
0    24597.0
1    27773.0
2     8517.0
3    24597.0
Name: Arrival_Time, dtype: float64

Answered By - not_speshal