r - Delete grouped rows from a dataframe -

i have dataframe 48503 rows. dataframe has multiple user ids can make multiple simulations or 1 , way tell looking @ time difference between simulations user did. if difference in time on 20 secs long assume guy simulations done in different sessions.

i want make new column session number of simulation each user

(id) (simulation number) (simulation-time-difference)  (session)           1                0:00:00.00                1          2                0:00:08.22                1          3                0:00:20.67                2          4                0:00:05.38                2  b         5                0:00:00.00                1  b         6                0:00:03.32                1  b         7                0:00:28.45                2 

here 1 method using ave , cumsum after extracting seconds:

# extract seconds df$seconds <- as.numeric(gsub("^0:00:([0-9]{2}\\.[0-9]{2})", "\\1", df$v3)) # calculate session number df$session <- ave((df$seconds > 20), df$v1, fun=cumsum) + 1 

this produces desired output.

df   v1 v2         v3 seconds session 1   1 0:00:00.00    0.00       1 2   2 0:00:08.22    8.22       1 3   3 0:00:20.67   20.67       2 4   4 0:00:05.38    5.38       2 5  b  5 0:00:00.00    0.00       1 6  b  6 0:00:03.32  1    3.32       1 7  b  7 0:00:28.45  2   28.45       2 

data

df <- read.table(text="         1     0:00:00.00                       2                0:00:08.22                       3                0:00:20.67                       4                0:00:05.38               b         5                0:00:00.00               b         6                0:00:03.32    b         7                0:00:28.45", as.is=true)