Regular expression to search for whole phrases and single words at the same time in str

I'm confused about regular expressions in str_detect(). Here is the setup:

library(stringr)

vector_to_check <- c("overall survival (os) in melanoma participants (parts b plus d)", "median overall survival (os)", "one- and two-year overall survival rate will be determined.", "overall survival rate (os rate) at month 6 in all participants")

str_detect(vector_to_check, "rate") 
# [1] FALSE FALSE  TRUE  TRUE

str_detect(vector_to_check , "overall survival (os) in melanoma participants (parts b plus d)")
# [1] FALSE FALSE FALSE FALSE

Basically, I want to input two types of pattern in str_detect(string, pattern):

Single words like "rate", "median", etc
Whole phrases, like "overall survival (os) in melanoma participants (parts b plus d)"

Is there a regular expression (pattern) that allows for this?

Thank you

CodePudding user response：

Wrap with fixed as there are metacharacters (()) in it, which may need to be either escaped (\\) otherwise

library(stringr)
str_detect(vector_to_check , fixed("overall survival (os) in melanoma participants (parts b plus d)"))
[1]  TRUE FALSE FALSE FALSE
str_detect(vector_to_check, fixed("rate")) 
[1] FALSE FALSE  TRUE  TRUE

If we need to combine both,

library(purrr)
map(c("rate", "overall survival (os) in melanoma participants (parts b plus d)"), 
  ~ str_detect(vector_to_check, fixed(.x))) %>%
    reduce(`|`)

-output

[1]  TRUE FALSE  TRUE  TRUE

CodePudding user response：

You can use the collapse argument of paste0() to add an 'or' to any amount of patterns, like this:

library(stringr)

vector_to_check <- c("overall survival (os) in melanoma participants (parts b plus d)", "median overall survival (os)", "one- and two-year overall survival rate will be determined.", "overall survival rate (os rate) at month 6 in all participants")

patterns <- c("rate", 
              "overall survival \\(os\\) in melanoma participants \\(parts b plus d\\)")

str_detect(vector_to_check, paste0(patterns, collapse = "|"))
[1]  TRUE FALSE  TRUE  TRUE

I also added the \\ in front of the parenthesis, as in regular expression you'll need to escape them!