I have a small problem with regex. I need to parse all words from string behind:
word - word2, word3, word4
Have tried to solve, but it is returning only last iteration
(\w+) - ((\w+)[, ]{0,2})+
https://regex101.com/r/2Uot2M/1
Thank you for any help.
P.S: I can't just match all words like (\w+). I need to match string with the format above.
If you are using PCRE regex library and you need to pre-validate a string before extracting words from it, you may use the following pattern:
(?:\G(?!^)|^(?=\w+ - (?:\w+[, ]{0,2})+$))\W*\K\w+
See the regex demo.
How it works
(?:\G(?!^)|^(?=\w+ - (?:\w+[, ]{0,2})+$)) - either the end of the previous match (\G(?!^)) or (|) start of a string (^) that is followed with the following pattern:
\w+ - 1+ word chars - - a hyphen enclosed with single spaces(?:\w+[, ]{0,2})+ - 1+ occurrences of:
\w+ - 1+ word chars[, ]{0,2} - 0 to 2 occurrences of a space or comma$ - end of string\W* - 0+ non-word chars\K - a match reset operator that discards all text matched so far from Group 0 (whole match) buffer\w+ - 1+ word chars.If you're just looking to capture each word into a separate capture group, you can just use this regex:
(\w+)
This captures all substrings with one or more word characters (letters or numbers). It will ignore the whitespace and punctuation. On Regex 101 it captures 'word', 'word2', 'word3', and 'word4' into separate capture groups
Helpful link on word characters: https://www.w3schools.com/jsref/jsref_regexp_wordchar.asp
Helpful link on quantifiers ('+' is a quantifier): https://learn.microsoft.com/en-us/dotnet/standard/base-types/quantifiers-in-regular-expressions
No validation needed except with first word-word.
Using the \G anchor and a branch reset, will fill an array
where the words collect in capture group 1.
(?|(\w+)[ ]*-[ ]*(?=\w)|(?!^)\G[ ]*,?[ ]*(\w+))
https://regex101.com/r/deZq5u/1
Note no need for BOS or EOS anchor's which are crutches.
This will find valid matches mid-string as it should.
(Note the # Optional spaces, single comma, spaces will always match
either a space or a comma or both, even though optional, and is a required
separator. This is because the \w+ clause will not leave any behind.)
(?| # Branch reset
( \w+ ) # (1), First word
[ ]* - [ ]* # qualified with a dash,
(?= \w ) # then a lookahead for next word
| # or,
(?! ^ ) # Reset \G at BOS
\G # Anchor, second or more match
[ ]* ,? [ ]* # Optional spaces, single comma, spaces
( \w+ ) # (1), Second or more word
) # End branch reset
$3captures. Or, a PCRE solution will look uglier,(?:\G(?!^)|^(?=\w+ - (?:\w+[, ]{0,2})+$))\W*\K\w+.