16

I am trying to get the number of open brackets in a character string in R. I am using the str_count function from the stringr package

s<- "(hi),(bye),(hi)"
str_count(s,"(")

Error in stri_count_regex(string, pattern, opts_regex = attr(pattern, : ` Incorrectly nested parentheses in regexp pattern. (U_REGEX_MISMATCHED_PAREN)

I am hoping to get 3 for this example

Sotos
  • 51,121
  • 6
  • 32
  • 66
NPHARD
  • 197
  • 1
  • 2
  • 5

3 Answers3

15

( is a special character. You need to escape it:

str_count(s,"\\(")
# [1] 3

Alternatively, given that you're using stringr, you can use the coll function:

str_count(s,coll("("))
# [1] 3
sebastian-c
  • 15,057
  • 3
  • 47
  • 93
5

You could also use gregexpr along with length in base R:

sum(gregexpr("(", s, fixed=TRUE)[[1]] > 0)
[1] 3

gregexpr takes in a character vector and returns a list with the starting positions of each match. I added fixed=TRUE in order to match literals.length will not work because gregexpr returns -1 when a subexpression is not found.

If you have a character vector of length greater than one, you would need to feed the result to sapply:

# new example
s<- c("(hi),(bye),(hi)", "this (that) other", "what")
sapply((gregexpr("(", s, fixed=TRUE)), function(i) sum(i > 0))
[1] 3 1 0
lmo
  • 37,904
  • 9
  • 56
  • 69
4

If you want to do it in base R you can split into a vector of individual characters and count the "(" directly (without representing it as a regular expression):

> s<- "(hi),(bye),(hi)"
> chars <- unlist(strsplit(s,""))
> length(chars[chars == "("])
[1] 3
John Coleman
  • 51,337
  • 7
  • 54
  • 119