删除字符串中的其他单词，除了需要的单词，包括 R 中的某些模式

Question

I have a vector including certain strings, and I would like remove other parts in each string except the word including certain patter (here is mir ).

s <- c("a mir-96 line (kk27)", "mir-133a cell", 
       "d mir-14-3p in", "m mir133 (sas)", "mir_23_5p r 27")

I want to obtain:

mir-96 , mir-133a , mir-14-3p , mir133 , mir_23_5p

I know the idea: use the gsub() and pattern is: a word beginning with (or including) mir .

But I have no idea how to construct such patter.

Or other idea?

Any help will be appreciated!

Answer 1

One way in base R would be splitting every string into words and then extracting only those with mir in it

unlist(lapply(strsplit(s, " "), function(x) grep("mir", x, value = TRUE)))
#[1] "mir-96"    "mir-133a"  "mir-14-3p" "mir133"    "mir_23_5p"

We can save the unlist step in lapply by using sapply as suggested by @Rich Scriven in comments

sapply(strsplit(s, " "), function(x) grep("mir", x, value = TRUE))

Answer 2

We can use sub to match zero or more characters ( .* ) followed by a word boundary ( \\\\b ) followed by the string ( mir and one or more characters that are not a white space ( \\\\S+ ), capture it as a group by placing inside the (...) followed by other characters, and in the replacement use the backreference of the captured group ( \\\\1 )

sub(".*\\b(mir\\S+).*", "\\1", s)
#[1] "mir-96"    "mir-133a"  "mir-14-3p" "mir133"    "mir_23_5p"

Update

If there are multiple 'mir.*' substring, then we want to extract strings having some numeric part

sub(".*\\b(mir[^0-9]*[0-9]+\\S*).*", "\\1", s1)
#[1] "mir-96"    "mir-133a"  "mir-14-3p" "mir133"    "mir_23_5p" "mir_23-5p"

data

s1 <- c("a mir-96 line (kk27)", "mir-133a cell", "d mir-14-3p in", "m mir133 (sas)", 
                                            "mir_23_5p r 27", "a mir_23-5p 1 mir-net")