带有 datetime64 [ns, UTC] 字段的 Pandas Timedelta 错误

Question

我有一个数据框，其中包含如下所示的两个时间字段。当我尝试在它们之间获取 Timedelta 时，我收到以下错误消息。我已经为这些字段包含了 .info() 。有谁知道问题是什么，你能建议如何解决它吗？非常感谢任何提示。

Data:

                            et_utc                    ts_utc
0 2018-05-02 09:24:29.304000+00:00 2018-05-02 09:39:15+00:00
1 2018-05-02 09:26:12.132000+00:00 2018-05-02 09:39:15+00:00
2 2018-05-02 09:28:37.913000+00:00 2018-05-02 09:39:12+00:00
3 2018-05-02 09:28:37.913000+00:00 2018-05-02 09:28:49+00:00
4 2018-05-02 10:39:48.820000+00:00 2018-05-02 10:39:48+00:00


Data description:

df[[‘et_utc','ts_utc']].info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 25625 entries, 0 to 25624
Data columns (total 2 columns):
et_utc    25625 non-null datetime64[ns, UTC]
ts_utc    25625 non-null datetime64[ns, UTC]
dtypes: datetime64[ns, UTC](2)
memory usage: 1.8 MB


Code:

df['t_delta']=pd.Timedelta(df['et_utc'] - df['ts_utc']).seconds


Error:

ValueError: Value must be Timedelta, string, integer, float, timedelta or convertible

Answer 1

如果你想以秒为单位获取时间增量，根据official documentation，你可以这样做：

df['t_delta']=(df['et_utc'] - df['ts_utc']).astype('timedelta64[s]')

Answer 2

使用Series dt accessor 访问日期时间（timedelta）系列的方法和属性。

>>> (df['et_utc'] - df['ts_utc']).dt.total_seconds()
0   -885.696
1   -782.868
2   -634.087
3    -11.087
4      0.820
dtype: float64
>>> df['t_delta'] = (df['et_utc'] - df['ts_utc']).dt.total_seconds()
>>>
>>> print(df)
                   et_utc              ts_utc  t_delta
0 2018-05-02 09:24:29.304 2018-05-02 09:39:15 -885.696
1 2018-05-02 09:26:12.132 2018-05-02 09:39:15 -782.868
2 2018-05-02 09:28:37.913 2018-05-02 09:39:12 -634.087
3 2018-05-02 09:28:37.913 2018-05-02 09:28:49  -11.087
4 2018-05-02 10:39:48.820 2018-05-02 10:39:48    0.820
>>> 

Datetimelike properties

---

astyping 相当于地板分割，它失去​​了一点分辨率。

>>> (df['et_utc'] - df['ts_utc']).astype('timedelta64[s]')
0   -886.0
1   -783.0
2   -635.0
3    -12.0
4      0.0
dtype: float64
>>>