python - pandas:按时间戳分组

Question

>>>> data_frame
                                                       flog  \
                                                      first   
0037B4F499705D725C2B3B00956B574E                        NaT   
009A27C31C21B5C47ADD76F1C0158864                        NaT   
016E73F1038CE46AF4A619453AC7DE70 2014-06-03 05:03:38.869000   
0250F3B15665E8B00F7D58CCA8C2C8F4 2014-06-03 05:04:38.869000   
0260FA375596B150DF8B4D7E3CA2D934 2014-06-03 05:04:44.015000

我需要按分钟分组。例如，它应该可以说 2014-06-03 05:04: 2, 2014-06-03 05:03: 1
我怎样才能在pandas中做到这一点？

更新:数据实际上是这样的:

>>>> data_frame
                                                       flog  \                      glog \
                                                      first                         first
0037B4F499705D725C2B3B00956B574E                        NaT    2014-06-03 05:03:38.869000
009A27C31C21B5C47ADD76F1C0158864                        NaT    2014-06-03 05:04:38.869000
016E73F1038CE46AF4A619453AC7DE70 2014-06-03 05:03:38.869000    2014-06-03 05:05:38.869000
0250F3B15665E8B00F7D58CCA8C2C8F4 2014-06-03 05:04:38.869000                           NaT
0260FA375596B150DF8B4D7E3CA2D934 2014-06-03 05:04:44.015000                           naT

如何显示:

TIME                 flog      glog 
                    first     first
2014-06-03 05:03:       1         1
2014-06-03 05:04:       2         1
2014-06-03 05:05:       0         1

Answer 1

只需重新采样。请注意，NaT 目前是一个错误(0.14.0 中)，因此您需要先删除它们。

In [18]: df.dropna().set_index('date').resample('1T',how='count')
Out[18]: 
                     value
date                      
2014-06-03 05:03:00      1
2014-06-03 05:04:00      2