我有id和pid's的以下时间序列数据框。我有兴趣找出每个id 0的暂停（id not 0）和没有动作（pid）。
这里我认为最小ID至少为3（id==0 >=3）
dateTime id pid
5/21/2016 15:23 3:23:25 0 a
5/21/2016 15:23 3:23:26 0 a
5/21/2016 15:23 3:23:27 0 a
5/21/2016 15:23 3:23:28 0 a
5/21/2016 15:23 3:23:29 0 a
5/21/2016 15:23 3:23:30 1 a
5/21/2016 15:23 3:23:31 2 a
5/21/2016 15:23 3:23:32 3 a
5/21/2016 15:23 3:23:33 0 a
5/21/2016 15:23 3:23:34 1 a
5/21/2016 15:23 3:40:25 0 a
5/21/2016 15:23 3:40:26 0 a
5/21/2016 15:23 3:40:27 0 a
5/21/2016 15:23 3:40:28 1 a
5/21/2016 15:23 3:40:29 1 a
5/21/2016 15:23 3:23:35 0 b
5/21/2016 15:23 3:23:36 0 b
5/21/2016 15:30 3:30:02 0 b
5/21/2016 15:30 3:30:03 2 b
5/21/2016 15:30 3:30:04 3 b
5/21/2016 15:30 3:30:05 1 b
5/21/2016 15:30 3:30:06 2 b
5/21/2016 15:30 3:30:07 0 b
我期待上述数据的以下输出
nr_paused nr_move pid
2 2 a
1 1 b
nr_paused = Occurrences of id '0' for more than 3 times then each packet should be counted as 1.
In the above example, there are two such instances (rows 1-5 and 11-13) hence nr_paused for id 'a' is 2
Similarly, for b nr_paused=1 (rows 16-18)
The count nr_move should be id > 0 rows