Question

I have a dataset that I’m trying to chunk up into "events" based on a condition. I want to create a consecutive group number (ID) which increases each time the condition is met.

Some kinds of records indicate that a new event has started, while other kinds of records represent no change / staying the course.

For example, in this dataset whenever ‘Action’ is "Left" or "Right", a new event has started and ‘Id’ should be incremented by 1:

|  Id | Action  |
|-----+---------|
|   1 | Left    |
|   2 | Forward |
|   3 | Forward |
|   4 | Right   |
|   5 | Forward |
|   6 | Left    |
| ... | ...     |

The resulting table I want would look like:

|  Id | Action  | GroupId |
|-----+---------+---------|
|   1 | Left    |       1 |
|   2 | Forward |       1 |
|   3 | Forward |       1 |
|   4 | Right   |       2 |
|   5 | Forward |       2 |
|   6 | Left    |       3 |
| ... | ...     |     ... |

In something like python I might do this with a counter and a for loop (pseudo-ish code):

GroupID = 1
for row in data:
    if Action == "Left" OR Action == "Right":
        GroupID = GroupID + 1
    else:
        GroupID = GroupID

I feel like this should be a really simple one-liner, but my brain is broken right now and I’m having a hard time conceptualizing this.

Answer 1

GroupId = cumsum(Action %in% c("Left", "Right"))