Search code examples
rrangeaggregate

Calculate min and max (range) by group


I have something like this in a data frame:

PersonId Date_Withdrawal
       A      2012-05-01   
       A      2012-06-01
       B      2012-05-01
       C      2012-05-01
       A      2012-07-01
       A      2012-10-01
       B      2012-08-01
       B      2012-12-01
       C      2012-07-01

I'd like to obtain the min and max date by 'PersonId'


Solution

  • First, convert to a proper date class (always a good practice) and then you could run a simple range by group. Here's an attempt

    library(data.table)
    setDT(df)[, Date_Withdrawal := as.IDate(Date_Withdrawal)]
    df[, as.list(range(Date_Withdrawal)), by = PersonId]
    #    PersonId         V1         V2
    # 1:        A 2012-05-01 2012-10-01
    # 2:        B 2012-05-01 2012-12-01
    # 3:        C 2012-05-01 2012-07-01
    

    Or

    library(dplyr)
    df %>%
      mutate(Date_Withdrawal = as.Date(Date_Withdrawal)) %>%
      group_by(PersonId) %>%
      summarise(Min = min(Date_Withdrawal), Max = max(Date_Withdrawal))
    # Source: local data frame [3 x 3]
    # 
    #  PersonId        Min        Max
    #    (fctr)     (date)     (date)
    # 1        A 2012-05-01 2012-10-01
    # 2        B 2012-05-01 2012-12-01
    # 3        C 2012-05-01 2012-07-01
    

    P.S. base aggregate would look like aggregate(as.Date(Date_Withdrawal) ~ PersonId, df, range) but it refuses to retain classes .