How to pass different arguments to each group in grouping of data.table?

Question

Example:

Here is a data table called dt:

> library(data.table)
> dt <- data.table(colA=rep(letters[1:3],each=3), colB=0:8)
> dt
   colA colB
1:    a    0
2:    a    1
3:    a    2
4:    b    3
5:    b    4
6:    b    5
7:    c    6
8:    c    7
9:    c    8

I want to know:

For colA equals "a", is there any values in colB > 2?

For colA equals "b", is there any values in colB > 3?

For colA equals "c", is there any values in colB > 4?

I create a vector called arg to hold arguments for group "a", "b" & "c":

arg <- c(2,3,4)

Could anyone give me a simple way to pass arg to grouping of dt by colA?

Here is my desired result:

     colA    V1
  1:    a FALSE
  2:    b  TRUE
  3:    c  TRUE

This is my first question here and I tried to make it simple. Thank you in advance.

Answer 1

For each subgroup that it operates on, [.data.table() stores information about the current value(s) of the grouping variable(s) in a variable named .BY.

If you first set up a named vector that maps the grouping variable's levels to the desired parameter values, you can use .BY to index into it, extracting the appropriate values, like so:

arg <- setNames(c(2, 3, 4), c("a", "b", "c"))
arg
# a b c 
# 2 3 4

dt[, any(colB > arg[unlist(.BY)]), by="colA"]
#    colA    V1
# 1:    a FALSE
# 2:    b  TRUE
# 3:    c  TRUE