Get all pairs by groups in data.table

Question

I want to extract all pairs within groups of a data.table by two columns.

I have the following code:

dt = data.table(g1 = c(1, 1, 1, 2, 2, 2, 2, 2),
                g2 = c(3, 3, 4, 5, 5, 6, 6, 6),
                values = letters[1:8])

dt[, .(ind = combn(values, 2, paste0, collapse = "")), by = .(g1, g2)]

But this returns the error

Error in combn(values, 2, paste0, collapse = "") : n < m

The ideal result is some object that contains groups such as:

a,b
c
d,e
f,g
f,h
g,h

If c was not included because there is only one value, that is not a huge concern to me.

Is this possible to get using data.table?

Thanks in advance.

Note: This answer does what I want for one level. Generate All ID Pairs, by group with data.table in R

Answer 1

You can try the code below

> setnames(dt[, combn(values, min(2, .N), toString), by = .(g1, g2)], "V1", "ind")[]
      g1    g2    ind
   <num> <num> <char>
1:     1     3   a, b
2:     1     4      c
3:     2     5   d, e
4:     2     6   f, g
5:     2     6   f, h
6:     2     6   g, h

or

> dt[, .(ind = combn(values, min(2, .N), toString, FALSE)), by = .(g1, g2)][, ind := unlist(ind)][]
      g1    g2    ind
   <num> <num> <char>
1:     1     3   a, b
2:     1     4      c
3:     2     5   d, e
4:     2     6   f, g
5:     2     6   f, h
6:     2     6   g, h

where min(2, .N) is to rectify the number of selected elements in case that it is less than 2.