CODEWITHSUNDEEP
rdata.table
User2321
User2321

Reputation: 3062

R data table group values into vector by group

I have the following data table:

library(data.table)
dat = data.table(j = c(3,8,9,11,10,28, 5), gr = c(9,9,9,9,10,10, 10), day = c(1,1,2,2,1,1,2)) 
  > dat
    j gr day
1:  3  9   1
2:  8  9   1
3:  9  9   2
4: 11  9   2
5: 10 10   1
6: 28 10   1
7:  5 10   2

I would like for each group (gr column) and day to create a vector with all the available values in j.

I have tried the following to no avail:

dat[, j[1:.N], by = .(gr, day )]
dat[, list(j[1:.N]), by = .(gr, day )]

This partially works but I would need to do some further manipulations to get the desired result:

dat[, list(list(Reduce(c, j, accumulate = T))), by = .(gr, day )]

My ideal result would look like:

> res_dat
   gr day all_values
1:  9   1        3,8
2:  9   2       9,11
3: 10   1      10,28
4: 10   2          5

I do not want this to be a string, I want it to be a list of vectors so this answer data table string concatenation of SD columns for by group values is not applicable.

Upvotes: 2

Views: 854

Answers (1)

akrun
akrun

Reputation: 886968

We need a list column. The .( is a concise syntax for list( in data.table.

dat[, .(j = .(j)), by = .(gr, day)]

-output

   gr day     j
1:  9   1   3,8
2:  9   2  9,11
3: 10   1 10,28
4: 10   2     5

i.e. it can be otherwise written as

dat[, list(j = list(j)), .(gr, day)]
   gr day     j
1:  9   1   3,8
2:  9   2  9,11
3: 10   1 10,28
4: 10   2     5

Upvotes: 4

Related Questions