Count number of unique activities

Question

I have a dataframe (simplified version) as follows:

df <- data.frame(ID = rep('A',7), Activity = 
 c('Login','Login','cat1','Login','cat2','Login','Login'))

  ID Activity
1  A    Login
2  A    Login
3  A     cat1
4  A    Login
5  A     cat2
6  A    Login
7  A    Login

Here is what I wish to do:

• start from the first row

• initiate session=0

• create a dataframe to hold path and count

• If the Activity is equal to Login, then session=1, check the next row's Activity and record it. This will be a path until the next Login

• continue until you hit the next Login, then set session=2.

• The final outcome for this example would be:

         Path          count
         Login           3
         Login, cat1     1
         Login, cat12    1
  

Answer 1

Create groups based on "Login", split, then paste per group, finally, aggregate using table:

data.frame(
  table(
    sapply(split(df$Activity, cumsum(df$Activity == "Login")), function(i){
      paste(i, collapse = ",")
    })))
#         Var1 Freq
# 1      Login    3
# 2 Login,cat1    1
# 3 Login,cat2    1