Question

I have a pandas DataFrame containing timestamped events. Each event has a start time and end time:

start  end other_vars
  100  120  ...
  150  151  ...
  160  170  ...
  200  210  ...

Is there a clean way to calculate time between events (e.g. the span between the end of the previous event and the start of this event) in pandas?

start  end between other_vars
  100  120      NA   ...
  150  151      30   ...
  160  170       9   ...
  200  210      30   ...

Answer 1

I think the easiest way to accomplish this is to subtract one shifted column from the other. The shift function does exactly that, it shifts an array by a default of one index.

In [3]: df
Out[3]:
   start  end
0    100  120
1    150  151
2    160  170
3    200  210

In [4]: df.start - df.end.shift()
Out[4]:
0   NaN
1    30
2     9
3    30

In [5]: df['elapsed'] = df.start - df.end.shift()

In [6]: df
Out[6]:
   start  end  elapsed
0    100  120      NaN
1    150  151       30
2    160  170        9
3    200  210       30