I only want to replace string occurrences that follow a particular keyword/pattern and not before. in other words, do nothing until the first occurrence of the keyword-pattern, and then start to gsub to the right of that keyword-pattern. See below:
gsub("\\[|\\]", "", "ab[ cd] ef keyword [ gh ]keyword ij ")
Actual results:
"ab cd ef keyword gh keyword ij "
Desired results:
"ab[ cd] [][asfg] ]] ef keyword gh keyword ij "
[Edited to fix the results. I don't want to remove 'keyword']
[Edited to show case of multiple occurrences of keyword]
You might use \G to get continous matches after keyword. Use \K to forget what was matched and match the following [ or ] to be replaced with an empty string.
(?:^.*?keyword\b|\G(?!^))[^\[\]]*\K[\[\]]
In parts
(?: Non capturing group
^.*?keyword Match until the first keyword
| Or
\G(?!^) Assert position at the end of previous match, not at the start to get continuous matches
) Close non capturing group
[^\[\]]*\K Match 0+ times not [ or ] and forget what was matched using \K
[\[\]] Match either [ or ]
Regex demo | R demo
Your code might look like
gsub("(?:^.*?keyword\\b|\\G(?!^))[^\\[\\]]*\\K[\\[\\]]", "", "ab[ cd] ef keyword [ gh ]keyword ij ", perl=T)
Note to use perl=T at the end for Perl-like regular expressions.
Related
I have seen in manuals how to use grep to match either a pattern or an exact string. However, I cannot figure out how to do both at the same time. I have a latex file where I want to find the following pattern:
\caption[SOME WORDS]
and replace it with:
\caption[\textit{SOME WORDS}]
I have tried with:
texfile <- sub('\\caption[','\\caption[\\textit ', texfile, fixed=TRUE)
but I do not know how to tell grep that there should be some text after the square bracket, and then a closed square bracket.
You can use
texfile <- "\\caption[SOME WORDS]" ## -> \caption[\textit{SOME WORDS}]
texfile <-gsub('(\\\\caption\\[)([^][]*)]','\\1\\\\textit{\\2}]', texfile)
cat(texfile)
## -> \caption[\textit{SOME WORDS}]
See the R demo online.
Details:
(\\caption\[) - Group 1 (\1 in the replacement pattern): a \caption[ string
([^][]*) - Group 2 (\2 in the replacement pattern): any zero or more chars other than [ and ]
] - a ] char.
Another solution based on a PCRE regex:
gsub('\\Q\\caption[\\E\\K([^][]*)]','\\\\textit{\\1}]', texfile, perl=TRUE)
See this R demo online. Details:
\Q - start "quoting", i.e. treating the patterns to the right as literal text
\caption[ - a literal fixed string
\E - stop quoting the pattern
\K - omit text matched so far
([^][]*) - Group 1 (\1): any zero or more non-bracket chars
] - a ] char.
I would like to use R to remove all underlines expect those between words. At the end the code removes underlines at the end or at the beginning of a word.
The result should be
'hello_world and hello_world'.
I want to use those pre-built classes. Right know I have learn to expect particular characters with following code but I don't know how to use the word boundary sequences.
test<-"hello_world and _hello_world_"
gsub("[^_[:^punct:]]", "", test, perl=T)
You can use
gsub("[^_[:^punct:]]|_+\\b|\\b_+", "", test, perl=TRUE)
See the regex demo
Details:
[^_[:^punct:]] - any punctuation except _
| - or
_+\b - one or more _ at the end of a word
| - or
\b_+ - one or more _ at the start of a word
One non-regex way is to split and use trimws by setting the whitespace argument to _, i.e.
paste(sapply(strsplit(test, ' '), function(i)trimws(i, whitespace = '_')), collapse = ' ')
#[1] "hello_world and hello_world"
We can remove all the underlying which has a word boundary on either of the end. We use positive lookahead and lookbehind regex to find such underlyings. To remove underlying at the start and end we use trimws.
test<-"hello_world and _hello_world_"
gsub("(?<=\\b)_|_(?=\\b)", "", trimws(test, whitespace = '_'), perl = TRUE)
#[1] "hello_world and hello_world"
You could use:
test <- "hello_world and _hello_world_"
output <- gsub("(?<![^\\W])_|_(?![^\\W])", "", test, perl=TRUE)
output
[1] "hello_world and hello_world"
Explanation of regex:
(?<![^\\W]) assert that what precedes is a non word character OR the start of the input
_ match an underscore to remove
| OR
_ match an underscore to remove, followed by
(?![^\\W]) assert that what follows is a non word character OR the end of the input
I would like to find the rows in a vector with the word 'RT' in it or 'R' but not if the word 'RT' is preceded by 'no'.
The word RT may be preceded by nothing, a space, a dot, etc.
With the regex, I tried :
grep("(?<=[no] )RT", aaa,ignore.case = FALSE, perl = T)
Which was giving me all the rows with "no RT".
and
grep("(?=[^no].*)RT",aaa , perl = T)
which was giving me all the rows containing 'RT' with and without 'no' at the beginning.
What is my mistake? I thought the ^ was giving everything but the character that follows it.
Example :
aaa = c("RT alone", "no RT", "CT/RT", "adj.RTx", "RT/CT", "lang, RT+","npo RT" )
(?<=[no] )RT matches any RT that is immediately preceded with "n " or "o ".
You should use a negative lookbehind,
"(?<!no )RT"
See the regex demo.
Or, if you need to check for a whole word no,
"(?<!\\bno )RT"
See this regex demo.
Here, (?<!no ) makes sure there is no no immediately to the left of the current location, and only then RT is consumed.
Say I have a dataframe df in which a column df$strings contains strings like
[cat 00.04;09]
[cat 00.04;10]
and so on. I want to remove all characters between "[cat" and "]" to yield
[cat]
[cat]
I've tried this using gsub but it's not working and I'm not sure what I'm doing wrong:
gsub('cat*?\\]', '', df)
Note that cat*?\\] patten matches ca, then any 0+ t chars but as few as possible and then ].
You want to match any chars other than ] between [cat and ]:
gsub('\\[cat[^]]*\\]', '[cat]', df$strings)
Here,
\\[ - matches [
cat - matches cat
[^]]* - 0+ chars other than ] (note that ] inside the bracket expression should not be escaped when placed at the start - else, if you escape it, you will need to add perl=TRUE argument since PCRE regex engine can handle regex escapes inside bracket expressions (not the default TRE))
\\] - a ] (you do not even need to escape it, you may just use ]).
See the R demo:
x <- c("[cat 00.04;09]", "[cat 00.04;10]")
gsub('\\[cat[^]]*\\]', '[cat]', x)
## => [1] "[cat]" "[cat]"
If cat can be any word, use
gsub('\\[(\\w+)[^]]*\\]', '[\\1]', x)
where (\\w+) is a capturing group with ID=1 that matches 1 or more word chars, and \\1 in the replacement pattern is a replacement backreference that stands for the group value.
I have a string format that I would like to select from a character vector. The form is
123 123 1234
where the two spaces can also be a hyphen. i.e. 3 digits followed by space or hyphen, followed by 3 digits, followed by space or hyphen, followed by 4 digits
I am trying to do this by the following:
grep("^([0-9]{3}[ -.])([0-9]{3}[ -.])([0-9]{4}$)",mytext)
however this yields:
integer(0)
What am I doing wrong?
Your string has a whitespace at the end, so you can either consider that white space, like so:
grep("^([0-9]{3}[ -.])([0-9]{3}[ -.])([0-9]{4} $)",mytext)
Or remove the end of line assertion "$", like so:
grep("^([0-9]{3}[ -.])([0-9]{3}[ -.])([0-9]{4})",mytext)
Also, as pointed out by Wiktor Stribiżew, the character class [ -.] will match any character in the range between " " and ".". To match "-","." and " " you have to escape the "-" or put it at the end of the class. Like [ \-.] or [ .-]