Regex to match a list of one or more comma-separated words, unless the string ends in a comma

Question

I have written the following regex long time ago that must match at least 3 words and works in both Latin and Cyrillic characters : regex = '([^ ,;\d]{2,}[ ,;]{1,}){2,}[^ ,;\d]{2,}'

I would like to rewrite it to match hello but fail to match hello, because of the comma. However, I would still like it to match hello, and, more, words.

Example matches: hello, hello, test69, hello, test69, matches

Example non-matches: hello, hello test69, hello test69 matches

It's really not clear what you want, or why you are not using the obvious and simple `[^,]+` — tripleee, Sep 16 '22 at 10:03
@tripleee because we need the comma to be there, otherwise "two words" would match and we want it to fail, only match if it were "two, words" ... etc — DarknessPlusPlus, Sep 16 '22 at 10:47
You can use the [regex module](https://stackoverflow.com/questions/1832893/python-regex-matching-unicode-properties) to specify the [Unicode category or script instead](https://www.regular-expressions.info/unicode.html), eg `\p{Letter}+`, `\p{Latin}`, `\p{Cyrillic}` — Panagiotis Kanavos, Sep 16 '22 at 11:02
Please review the [help] and in particular [How to ask](/help/how-to-ask) as well as the guidance for providing a [mre]. — tripleee, Sep 16 '22 at 11:14
At least 3 whitespace/comma/semicolon-separated words containing 2 or more chars: `regex = r'\w{2,}(?:[\s,;]+\w{2,}){2,}'`. Does it solve the problem? — Wiktor Stribiżew, Sep 16 '22 at 11:20
@WiktorStribiżew the idea is to match any number of words that are comma separated, but fail if there is no comma between the words — DarknessPlusPlus, Sep 16 '22 at 11:22
This would match both words that are and are not comma separated, and will fail for Cyrillic. — DarknessPlusPlus, Sep 16 '22 at 11:25
It will not fail for Cyrillic, `\w` matches any Unicode letters. Next, if you need to match at least two words, use `+` instead of `*` at the end, `r'\w+(?:\s*,\s*\w+)+'` — Wiktor Stribiżew, Sep 16 '22 at 11:36
`[^,]+(,[^,]+)+` matches strings with one or more commas. It is still quite unclear what you are actually trying to accomplish. — tripleee, Sep 16 '22 at 11:42
@tripleee I want to match examples "one", "one, two", and fail to match examples "one," , "one two", "one, two," and etc. Your solution will fail at "one" but match at "one,". — DarknessPlusPlus, Sep 16 '22 at 11:46
So, [look here](https://regex101.com/r/0LbFdw/1), does `^\w+(?:, *\w+)*$` work now? — Wiktor Stribiżew, Sep 16 '22 at 14:29
@WiktorStribiżew I would say that does work for all of OP's stated requirements, yes. — TylerH, Sep 16 '22 at 14:45
@TylerH Unless OP uses Python 2. Only in Python 2, `\w` does not match all Unicode letters. — Wiktor Stribiżew, Sep 16 '22 at 14:46
@WiktorStribiżew that works great, do post is as an answer for future reference — DarknessPlusPlus, Sep 16 '22 at 14:50

score 2 · Accepted Answer · answered Sep 16 '22 at 14:52

You can use

^\w+(?:, *\w+)*$

In Python, you can use a shorter version if you use re.fullmatch:

re.fullmatch(r'\w+(?:, *\w+)*', text)

See the regex demo.

Note that in case your spaces can be any whitespaces, replace the with \s in the regex. If your words can only contain letters, replace each \w with [^\W\d_]. If your words can only contain letters and digits, replace every \w with [^\W_].

Details:

^ - start of string
\w+ - one or more word chars
(?:, *\w+)* - zero or more repetitions of a comma, zero or more spaces, and then one or more word chars
$ - end of string.

Regex to match a list of one or more comma-separated words, unless the string ends in a comma

1 Answers1