Question

I need to go from this

 id  |    date
-----------------
  A  | 2000-01-13
  A  | 2000-01-18
  A  | 2000-01-25
  B  | 2012-10-10
  B  | 2012-10-11
  C  | 2005-07-25
  C  | 2005-07-31

to this

 id  |    date     | days from start
---------------------------
  A  | 2000-01-13  |  0
  A  | 2000-01-18  |  5
  A  | 2000-01-25  |  12
  A  | 2000-02-08  |  26
  B  | 2012-10-10  |  0
  B  | 2012-10-11  |  1
  C  | 2005-07-25  |  0
  C  | 2005-07-31  |  6

i.e. creating a variable that holds the number of days passed since the first date, grouped by id.

Any ideas?

Answer 1

Using data.table: (I assume the date column is character here. If its date format, then you can remove the as.Date(.) function call.

df <- structure(list(id = c("A", "A", "A", "B", "B", "C", "C"), 
             date = c("2000-01-13", "2000-01-18", "2000-01-25", "2012-10-10", 
                    "2012-10-11", "2005-07-25", "2005-07-31")), 
             .Names = c("id", "date"), row.names = c(NA, -7L), 
             class = "data.frame")
require(data.table)
dt <- data.table(df, key="id")
dt[, days_from_start := cumsum(c(0, diff(as.Date(date)))),by=id]

#    id       date days_from_start
# 1:  A 2000-01-13               0
# 2:  A 2000-01-18               5
# 3:  A 2000-01-25              12
# 4:  B 2012-10-10               0
# 5:  B 2012-10-11               1
# 6:  C 2005-07-25               0
# 7:  C 2005-07-31               6