如何在组之间进行时间差异

Question

I have a problem related with timing differences and I am trying to solve via dplyr . My initial data frame looks like this :

Paper <- data.frame(
  Student = c("A", "A", "A", "A", "A", "A", "B", "B", "B", "B"), 
  Dates = c("2014-04-17", "2014-04-17", "2014-04-17", "2014-04-17", "2014-04-18", "2014-04-18", "2014-04-18", "2014-04-18", "2014-04-18","2014-04-18"),
  Time = c("10:35:00", "11:25:00", "19:15:00", "21:00:00", "22:00:00", "22:21:26", "10:25:00", "11:15:00", "16:05:00", "17:25:00"),
  Connection = c("Initial", "Final", "Initial", "Final", "Initial", "Final", "Initial", "Final", "Initial", "Final")
)

or

   Student      Dates     Time  Connection
       A    2014-04-17  10:35:00    Initial
       A    2014-04-17  11:25:00      Final
       A    2014-04-17  19:15:00    Initial
       A    2014-04-17  21:00:00      Final
       A    2014-04-18  22:00:00    Initial
       A    2014-04-18  22:21:26      Final
       B    2014-04-18  10:25:00    Initial
       B    2014-04-18  11:15:00      Final
       B    2014-04-18  16:05:00    Initial
       B    2014-04-18  17:25:00      Final

I am trying to to know for each Date the time dedicated by Student considering that the real time calculated is between the "Initial" and "Final" Connection .

So my expected data frame would look like this :

  Student    Dates    Time (Minutes)
     A    14-04-17     155
     A    14-04-18   21.43
     B    14-04-18     130

I have tried this, and I almost got the solution but I don't know how to consider the calculation of the difference of time between connection ( "Initial" / "Final" ) so I obtain this:

Paper$Dates <- as.Date(Paper$Dates, "%Y-%m-%d")

Paper$Time <- as.numeric(as.POSIXct(as.character(Paper$Time),
                         format = "%H:%M:%S"))

FinalPaper <- 
  Paper %>% 
  group_by(Student, Dates) %>% 
  summarise(TimeSpent = sum(diff(Time))) %>% 
  mutate(TimeSpent = TimeSpent/60) %>% 
  mutate(TimeSpent = round(TimeSpent, digits = 2))

Resulting

  Student      Dates   TimeSpent
1       A   2014-04-17    625.00
2       A   2014-04-18     21.43
3       B   2014-04-18    420.00

As can be seen in the TimeSpent the time is higher this is because I am not considering the connection so it is calculating wrong times. For example for the student A it is calculating the time between 10:35:00 and 21:00:00 which is wrong.

Thank you very much!!

Answer 1

You could add an id to each 'session', with cumsum(Connection == "Initial") . Prerequisite for this is that the data is sorted in the way you have presented it here. We can then calculate the time difference for each session, and aggregate again to get the total time spent per student per date:

Paper <- data.frame(
  Student = c("A", "A", "A", "A", "A", "A", "B", "B", "B", "B"), 
  Dates= c("2014-04-17", "2014-04-17", "2014-04-17", "2014-04-17", "2014-04-18", "2014-04-18", "2014-04-18", "2014-04-18", "2014-04-18","2014-04-18"),
  Time = c("10:35:00", "11:25:00", "19:15:00", "21:00:00", "22:00:00", "22:21:26", "10:25:00", "11:15:00", "16:05:00", "17:25:00"), 
  Connection = c("Initial", "Final", "Initial", "Final", "Initial", "Final", "Initial", "Final", "Initial", "Final")
)

Paper$Dates <- as.Date(Paper$Dates, "%Y-%m-%d")
Paper$Time <- as.numeric(as.POSIXct(as.character(Paper$Time),
                                    format = "%H:%M:%S"))

FinalPaper <- Paper %>% 
  mutate(seqid = cumsum(Connection == "Initial")) %>% 
  group_by(Student, Dates, seqid) %>% 
  summarise(TimeSpent = sum(diff(Time))) %>% 
  group_by(Student, Dates) %>% 
  summarise(TimeSpent = round(sum(TimeSpent)/60,2))

Output:

# A tibble: 3 x 3
# Groups:   Student [2]
  Student      Dates TimeSpent
   <fctr>     <date>     <dbl>
1       A 2014-04-17    155.00
2       A 2014-04-18     21.43
3       B 2014-04-18    130.00

Hope this helps!

Answer 2

And here is a data.table based solution:

library(data.table)
setDT(Paper)
Paper[order(Student, Time), .(
    TimeSpend = sum(c(0,diff(Time))[Connection == "Final"])/60
  ), by = .(Student, Dates)]

   Student      Dates TimeSpend
1:       A 2014-04-17 155.00000
2:       A 2014-04-18  21.43333
3:       B 2014-04-18 130.00000