How can I perform a “setdiff” merge using data.table?

Question

Say I have two tables:

library(data.table)
set.seed(1)

tab1 <- data.table(
  let = rep(letters[1:2], each = 3),
  num = rep(1:3, 2),
  val = rnorm(6),
           


        

Answer 1

In this case, it's equivalent to an anti join:

tab1[!tab2, on=c("let", "num")]

But setdiff() would only the first row for every let,num. This is marked for v1.9.8, FR #547.

Answer 2

One solution would be to do a merge and remove the rows where there are values from tab2

d<-as.data.frame(merge(tab1,tab2,all=T))
t<-is.na(d[,4])
d[t,][,-4]

 let num      val.x
3   a   3 -0.8356286
6   b   3 -0.8204684

Using data.table :

merge(tab1,tab2,all=T)[is.na(val.y), 1:3]

 let num      val.x
1:   a   3 -0.8356286
2:   b   3 -0.8204684