Split a string without considering special characters

I need a way to split a string every n letters.

For example, let s="QW%ERT%ZU%I%O%P" and n=3, I want to obtain "QW%E" "RT%Z" "U%I%O" "%P".

As you can see, the special character "%" is not considered in the division.

I tried with

strsplit(s, "(?<=.{10})(?=.*\\%)", perl = TRUE)[[1]]

but I cannot find a way to obtain what I want.

What about regmatches (instead of strsplit) like below?

> n <- 3

> regmatches(s, gregexpr(sprintf("(\\W?\\w){1,%i}", n), s))
[[1]]
[1] "QW%E"  "RT%Z"  "U%I%O" "%P"

Or tapply + strsplit

v <- unlist(strsplit(s, ""))
l <- which(grepl("\\w", v))
tapply(
    v,
    cumsum(seq_along(v) %in% (1 + by(l, ceiling(seq_along(l) / n), max))),
    paste0,
    collapse = ""
)

which gives

      0       1       2       3
 "QW%E"  "RT%Z" "U%I%O"    "%P"

Much less succinct then the above but a Base Solution all the same:

# Function to only consider certain characters define by a regex
# and split a string scalar into seperate elements in a vector
split_string_to_vec <- function(s, n, consider_elements_pattern = "[[:alpha:]]"){
# Ensure s is a character scalar:
stopifnot(is.character(s) && length(s) == 1)
# Ensure n is an integer scalar: 
stopifnot(is.numeric(n) && length(n) == 1)
# Split the string into separate elements:
# str_vec => character vector
str_vec <- unlist(strsplit(s, ""))
# Assign an index to the string vector: 
# idx => named integer vector
idx <- setNames(seq_len(length(str_vec)), str_vec)
# Resolve which values are to be considered (only alpha numerics):
# considered_vals => named integer vector
considered_vals <- idx[grepl(consider_elements_pattern, names(idx))]
# Split the string vector into a list: 
# grpd_strings => list of character vectors
grpd_strings <- split(
considered_vals,
ceiling(seq_along(considered_vals) / n)
)
# For each string group, resolve the group with the 
# appropriate characters in order: res_vec => character vector
res_vec <- vapply(
seq_along(grpd_strings),
function(i){
# Get current list element: 
curr <- grpd_strings[[i]]
# If its the first element: 
if(i == 1){
# Ignore previous element only focus on this 
# one: ir => named integer vector
ir <- sort(c(curr, idx[min(curr):max(curr)]))
# Otherwise:
}else{
# Resolve the previous element: 
prev <- grpd_strings[[(i-1)]]
# ir => named integer vector
ir <- sort(c(curr, idx[(max(prev)+1):max(curr)]))
}
# Flatten result into a unique (by idx) string: 
# character scalar => env
paste0(
names(
subset(
ir,
!(duplicated(ir))
)
),
collapse = ""
)
},
character(1)
)
# Explicitly define the returned object:
# character vector => env
return(res_vec) 
}
# Input Data:
# s => string scalar
s <- "QW%ERT%ZU%I%O%P"
# n => integer scalar
n <- 3
# Apply the function: string scalar => stdout(console)
split_string_to_vec(s, n, consider_elements_pattern = "[[:alpha:]]")