extract string in R using regex for a variable number of instances

Question

I've got a string column (data.table) that I need to parse based on a pattern (text between '-' ) and a defined (but variable) number of instances of that pattern and I'm not sure how to do it using regex:

> test <- c("AAA-bb-ccc", "abcd-efgh","blah", "blah-blah-blah-blah")

say, the predefined number of instances is i.

> i = 1
> output
"AAA"  "abcd"  "blah"  "blah

> i = 2
> output
"bb"  "efgh"  ""  "blah"


> i= 3
> output
"ccc"  ""  ""  "blah"

how would i use a general regex using i that would achieve this?

Answer 1

We can create a function which splits on "-" and returns the ith value.

get_i_th_element <- function(test, i) {
   sapply(strsplit(test, "-"), function(x) if(length(x) >= i) x[[i]] else "")
}

get_i_th_element(test, 1)
#[1] "AAA"  "abcd" "blah" "blah"

get_i_th_element(test, 3)
#[1] "ccc"  ""     ""     "blah"

Answer 2

For i=3 you can try

unlist(lapply(strsplit(test,split = '-'),'[',3)) 
[1] "ccc"  NA     NA     "blah"

Answer 3

We can also use tokenize_regex from the tokenizers package and then data.table::transpose and cbind relevant columns into a data.table

test <- c("AAA-bb-ccc", "abcd-efgh","blah", "blah-blah-blah-blah")

library(tokenizers)
library(data.table)
test <- transpose(tokenize_regex(test, "-"), fill = "")

i <- 1:3
as.data.table(do.call(cbind, test[i]))
#     V1   V2   V3
#1:  AAA   bb  ccc
#2: abcd efgh
#3: blah
#4: blah blah blah