Is there an optimal way to create a bunch of new columns with data.table?

Question

I am using data.table I have a vector of the names for the new columns. I want to create these new columns but using the information from the old columns. Let me show it on a following example:

data <- data.table(a = c("OneA", "TwoB", "ThreeC"),
                   b = c(1, 2, 3))

newCols <- c("One", "Two", "Three")

for (newCol in newCols) {
  data[, eval(newCol) := gsub(paste0("^.*", newCol), "", a)]
}

As an output I expect (and get) the following:

> data
        a   b    One    Two Three
1:   OneA   1      A   OneA  OneA
2:   TwoB   2   TwoB      B  TwoB
3: ThreeC   3 ThreeC ThreeC     C

In this case I combine defining new columns from the vector and using the vector values themselves to fill these columns. Is there a way to do it more optimal (for example with a set())?

Answer 1

One possible approach could be:

library(data.table)

DT[, (newCols) := lapply(newCols, function(x) sub(x, "", a))][]
#>         a b    One    Two Three
#> 1:   OneA 1      A   OneA  OneA
#> 2:   TwoB 2   TwoB      B  TwoB
#> 3: ThreeC 3 ThreeC ThreeC     C

Data

DT <- data.table(a = c("OneA", "TwoB", "ThreeC"), b = c(1, 2, 3))
newCols <- c("One", "Two", "Three")