Regex to only match words without special symbols or numbers

Question

I'm trying to figure out how to match "word" but not "word1", "w1ord" or any variation of "word!" where "!" could be anything "#%!"

Basically I want to match against word that only uses A-Z and a-z. The following does not work :/

([A-Za-z])\w+

As "word%5" is a match.

Well, now, I think you need `Regex.Matches(str, @"(?<!\S)[a-zA-Z]+(?!\S)")` — Wiktor Stribiżew, Nov 15 '16 at 09:02
So, are you trying to check a word or *string* for English only letters? — Wiktor Stribiżew, Nov 15 '16 at 09:36

mybirthname · Accepted Answer · 2016-11-15T09:25:43.853

1

You can do it with linq. Look the ASCII Table to see the letter values from A to Z for chars. You can Upper your string after that only to check from range 65 to 90.

bool notOnlyLetters = yourStringValue.ToUpper().Any(x => !(x >= 65 && x <= 90));

edited Nov 15 '16 at 09:25

answered Nov 15 '16 at 08:58

mybirthname

17,949
3
31
55

1

This will find non-English letters, too. – Wiktor Stribiżew Nov 15 '16 at 08:58
@WiktorStribiżew ok you are right I miss that he want only from A to Z. – mybirthname Nov 15 '16 at 08:59
@WiktorStribiżew now should be okay – mybirthname Nov 15 '16 at 09:03
I do not think it is working. You included some non-letters into the range. See http://ideone.com/IATcQx – Wiktor Stribiżew Nov 15 '16 at 09:07
@WiktorStribiżew okay again right I didn't look carefully the ASCII Table. Now it should be okay. – mybirthname Nov 15 '16 at 09:21
Note that missing explanations in the answers usually lead to "clarification" questions in the future. Besides, there is no point using `ToArray` as a string is already treated as a char array with Linq. – Wiktor Stribiżew Nov 15 '16 at 09:23
@WiktorStribiżew Point taken I wrote little explanation and add link to ASCII Table, thanks for the comments. – mybirthname Nov 15 '16 at 09:26
In some specific regional cultures, this `ToUpper()` solution still won't work in 100% cases. – Wiktor Stribiżew Nov 15 '16 at 09:34

Wiktor Stribiżew · Answer 2 · 2016-11-15T09:39:39.300

To match a whole string consisting of English letters, use LINQ or regex:

var hasAllEnglishLetters = x.All(c => (c >= 65 && c <= 90) || c >=97 && c<= 122));
var hasAllEnglishLetters = Regex.IsMatch(x, @"^[a-zA-Z]+$");

To match words inside a larger string, you may either use a regex or LINQ approaches, too:

var s = "Match word but not word1, w1ord or word!";
var res_linq = s.Split().Where(x => x.All(c => (c >= 65 && c <= 90) || c >=97 && c<= 122));
Console.WriteLine(string.Join(";", res_linq));
// REGEX
var res_regex = Regex.Matches(s, @"(?<!\S)[a-zA-Z]+(?!\S)").Cast<Match>().Select(m=>m.Value);
Console.WriteLine(string.Join(";", res_regex));

See the online C# demo

LINQ approach details: With Split(), the string is split into chunks of non-whitespace symbols and .All(c => (c >= 65 && c <= 90) || c >=97 && c<= 122) makes sure only those chunks are fetched that belong to ASCII letters (65 to 90 - uppercase ASCII letters, and 97 to 122 are lowercase ones).

Regex approach: the (?<!\S) lookbehind fails the match if there is no whitespace before [a-zA-Z]+ (or start of string), 1 or more ASCII letters, and the negative lookahead (?!\S) fails the match if there is no whitespace (or end of string) after the letters.

score 1 · Answer 3 · answered Nov 15 '16 at 09:27

The following regex accepts every "word" occurrence contained between:

the start of a line OR an empty space (^|\s)

and

an empty space OR the end of a line. (\s|$)

(^|\s)word(\s|$)

If you want to find every word composed by alphabetic characters only, you can change the regex as follows:

(^|\s)[a-zA-Z]+(\s|$)

Regex to only match words without special symbols or numbers

3 Answers3

Linked