Assign unique IDs by group, according to pattern

Question

I have a grouped dataset, but the groups don\'t have a unique identifier, like this:

direction <- c(\'N\',\'S\',\'W\',\'N\',\'N\',\'S\',\'W\',\'N\',\'S\',\'W\         


        

Answer 1

You can get this result with == to identify elements with "N" and cumsum to construct indices. Then pull out the values from the stored vector LETTERS as suggested in balter's answer.

Here, cumsum sums over a logical vector, coercing it to numeric binary (1s, and 0s). It thus repeats the same value and increments any time an "N" was encountered.

 x$ID <- LETTERS[cumsum(x$direction == "N")]

This returns

x
   direction measurement ID
1          N           4  A
2          S           6  A
3          W           1  A
4          N           7  B
5          N           2  C
6          S           4  C
7          W           7  C
8          N           4  D
9          S           1  D
10         W           4  D