For the
repeat type sensor data
, we use the information of start and end flag.
But we do not have such data for
fluent type sensor data.
For the fluent type, we use the knowledge of Time Difference of Cause-Effect. We can use the time of effect as the end of the time.
Start time is difficult. We need consider the maximum time of cause and effect.
In the real we search from recent data to make aggregated data. So I call "reverse time aggregation". The flags and direct of time are not different from repeat type. But others are same.