CODEWITHSUNDEEP
rregexstringgsub
Mark
Mark

Reputation: 1769

Use gsub to remove pattern from a string: argument 'pattern' has length > 1 and only the first element will be used

I am trying to remove pattern from string with gsub from the following array of chr

articles<-c("RT @name1: hello world", "@nickname1: bye bye guys", 
"RT @name2_surname2: I have no text", "Hello!")

The pattern is formed by the terms between @ and : only in the strings that begin with RT. Hence in our case the pattern is:

 "name1"    "name2_surname2"

The pattern can be obtained by using

pat <- "^RT.*?@(.*?):.*"
res <- gsub(pat,"\\1",articles[grepl(pat,articles)])

After removal of this pattern, the desired result is so:

"RT : hello world", "@nickname1: bye bye guys", 
"RT : I have no text", "Hello!"

Anyway, when I use:

gsub(res,"",articles)

I obtain a wrong result:

[1] "RT @: hello world"                  "@nick: bye bye guys"               
[3] "RT @name2_surname2: I have no text" "Hello!"                            
Warning message:
In gsub(res, "", articles) :
  argument 'pattern' has length > 1 and only the first element will be used

Upvotes: 1

Views: 474

Answers (2)

Chris Ruehlemann
Chris Ruehlemann

Reputation: 21400

If the desired output is, as stated, this:

"RT : hello world", "@nickname1: bye bye guys", "RT : I have no text", "Hello!"

then this solution works:

First, you need to change the pattern to include @ in the capturing group:

pat <- "^RT.*?(@.*?):.*"
res <- gsub(pat,"\\1",articles[grepl(pat,articles)])

Then, as suggested by @Akrun, you can paste the two vector elements of restogether, which allows you to use it as a (single) pattern:

gsub(paste0(res, collapse = "|"), "", articles)

That will give you the disired output.

Upvotes: 2

akrun
akrun

Reputation: 887028

We can paste the patterns to a single string and use that in gsub pattern as the pattern argument is not vectorized i.e, it takes only a length of 1

gsub(paste0("\\b(", paste(res, collapse="|"),  ")\\b"), "", articles)
#[1] "RT @: hello world"        "@nickname1: bye bye guys" "RT @: I have no text"     "Hello!"

Upvotes: 0

Related Questions