Pandas 中的日期时间转换问题

Question

Here is a reproduced example of my problem :

df = pd.DataFrame(["2018-01-13 17:25:54+0100",
    "2018-01-13 07:23:36+0100",
    "2018-01-13 08:15:48+0100"], columns=["date"])

print(type(datetime.strptime(df["date"][1], '%Y-%m-%d %H:%M:%S%z')))

print(type(pd.Series(df["date"].apply(lambda s: datetime.strptime(s, '%Y-%m-%d %H:%M:%S%z')))[1]))

The output are :

class 'datetime.datetime'

class 'pandas._libs.tslibs.timestamps.Timestamp'

How can I get a datetime.datetime object by using the apply function (or a similar one)?

Answer 1

Use Pandas methods for datetime operations

In general, avoid the datetime module from the standard library when dealing with Pandas dataframes. You should want to use vectorised operations, and should rely on Pandas methods taking advantage of NumPy-based vectorisation:

df['date'] = pd.to_datetime(df['date'])

print(df['date'].dtype)
# datetime64[ns]

But if you insist...

If you wish to export to an array of datetime.datetime values for use outside of Pandas, you can use to_pydatetime :

py_date = df['date'].dt.to_pydatetime()
# array([datetime.datetime(2018, 1, 13, 16, 25, 54),
#        datetime.datetime(2018, 1, 13, 6, 23, 36),
#        datetime.datetime(2018, 1, 13, 7, 15, 48)], dtype=object)

However, once you are using Pandas there's rarely a need to do so.