Regex to match only letters

Question

How can I write a regex that matches only letters?

Solution 1

Use a character set: [a-zA-Z] matches one letter from A–Z in lowercase and uppercase. [a-zA-Z]+ matches one or more letters and ^[a-zA-Z]+$ matches only strings that consist of one or more letters only (^ and $ mark the begin and end of a string respectively).

If you want to match other letters than A–Z, you can either add them to the character set: [a-zA-ZäöüßÄÖÜ]. Or you use predefined character classes like the Unicode character property class \p{L} that describes the Unicode characters that are letters.

Solution 2

\p{L} matches anything that is a Unicode letter if you're interested in alphabets beyond the Latin one

Solution 3

Depending on your meaning of "character":

[A-Za-z] - all letters (uppercase and lowercase)

[^0-9] - all non-digit characters

Solution 4

The closest option available is

[\u\l]+

which matches a sequence of uppercase and lowercase letters. However, it is not supported by all editors/languages, so it is probably safer to use

[a-zA-Z]+

as other users suggest

Solution 5

You would use

/[a-z]/gi

[]--checks for any characters between given inputs

a-z---covers the entire alphabet

g-----globally throughout the whole string

i-----getting upper and lowercase

View more solutions

Comments

[Philip Potter]

not in all regex flavours. For example, vim regexes treat \p as "Printable character".

[Joachim Sauer]

That's a very ASCII-centric solution. This will break on pretty much any non-english text.

[Philip Potter]

this page suggests only java, .net, perl, jgsoft, XML and XPath regexes support \p{L}. But major omissions: python and ruby (though python has the regex module).

[Gumbo]

@Joachim Sauer: It will rather break on languages using non-latin characters.

[Nike]

I meant lettters. It doesn't appear to be working though. preg_match('/[a-zA-Z]+/', $name);

[Ivo Wetzel]

Already breaks on 90% of German text, don't even mention French or Spanish. Italian might still do pretty well though.

[Joachim Sauer]

that depends on what definition of "latin character" you choose. J, U, Ö, Ä can all be argued to be latin characters or not, based on your definition. But they are all used in languages that use the "latin alphabet" for writing.

[KristofMols]

[A-Za-z] is just the declaration of characters you can use. You still need to declare howmany times this declaration has to be used: [A-Za-z]{1,2} (to match 1 or 2 letters) or [A-Za-z]{1,*} (to match 1 or more letters)

[Jörg W Mittag]

@Philip Potter: Ruby supports Unicode character properties using that exact same syntax.

[Amal Murali]

\w may not be a good solution in all cases. At least in PCRE, \w can match other characters as well. Quoting the PHP manual: "A "word" character is any letter or digit or the underscore character, that is, any character which can be part of a Perl "word". The definition of letters and digits is controlled by PCRE's character tables, and may vary if locale-specific matching is taking place. For example, in the "fr" (French) locale, some character codes greater than 128 are used for accented letters, and these are matched by \w.".

[OGHaza]

That is not what [^\W|\d] means

[OGHaza]

[^\W|\d] means not \W and not | and not \d. It has the same net effect since | is part of \W but the | does not work as you think it does. Even then that means it accepts the _ character. You are probably looking for [^\W\d_]

[Motlab]

I agree with you, it accepts the _. But "NOT" | is equal than "AND", so [^\W|\d] means : NOT \W AND NOT \d

[OGHaza]

[^ab] means not a and not b. [^a|b] means not a and not | and not b. To give a second example [a|b|c|d] is exactly the same as [abcd|||] which is exactly the same as [abcd|] - all of which equate to ([a]|[b]|[c]|[d]|[|]) the | is a literal character, not an OR operator. The OR operator is implied between each character in a character class, putting an actual | means you want the class to accept the | (pipe) character.

[V-SHY]

words include other characters from letters

[Nyerguds]

Won't match any special characters though.

[Eugen Konkov]

\w means match letters and numbers

[ZoFreX]

I think this should be \p{L}\p{M}*+ to cover letters made up of multiple codepoints, e.g. a letter followed by accent marks. As per regular-expressions.info/unicode.html

[phuclv]

well à, á, ã, Ö, Ä... are letters too, so are অ, আ, ই, ঈ, Є, Ж, З, ﺡ, ﺥ, ﺩא, ב, ג, ש, ת, ... en.wikipedia.org/wiki/Letter_%28alphabet%29

[Radu Simionescu]

\p{L} matches all the umlauts sedilla accents etc, so you should go with that.

[user1329482]

Works well in a selector engine for determining if the selector is just a tag name.

[DaveMongoose]

This will also match whitespace, symbols, etc. which does not seem to be what the question is asking for.

[AER]

What do you do if you can't use [] because Python is too thick to understand nestings?

[The Witness]

And what about for instance, “Zażółć gęslą jaźń”?

[karoluS]

it doesn't include diacritic signs such as ŹŻŚĄ

[matanster]

with python 3 this yields an error bad escape \p at position 0

[Pablo]

Instead of keep adding characters like adding äöüßÄÖÜ, you can go: ^[a-zA-Z]\p{L}+$ to include most of the western alphabets.

[Catalina Chircu]

@phuclv: Indeed, but that depends on the encoding, and the encoding is part of the settings of the program (either the default config or the one declared in a config file of the program). When I worked on different languages, I used to store that in a constant, in a config file.

[phuclv]

@CatalinaChircu encoding is absolutely irrelevant here. Encoding is a way to encode a code point in a character set in binary, for example UTF-8 is an encoding for Unicode. Letters OTOH depends on the language, and if one says [A-Za-z] are letters then the language that's being used must be specified

[Catalina Chircu]

@phuclv: Indeed, I should have mentioned the language, not the encoding. The language is important and finding the letters in English is not the same as finding the letters in Spanish or French. If you do not take into account the diacritics in these languages you can cut words in two.

[Stefan Haustein]

Doesn't work in firefox: bugzilla.mozilla.org/show_bug.cgi?id=1361876

[Toto]

You should have look at an ASCII table. A-z matches more than just letters, as well as À-ú

[ndrwnaguib]

Hello @jarraga. Welcome to SO, did you read how to answer a question?. It should assist the clearance of your answer, and hence avoid down voting.

[Toto]

What about non Latin letter? For example çéàñ. Your regex is less readable than \p{L}

[Frederic]

Clever answer. Works perfectly for accented letters as well.

[jave.web]

For letters beyond english: /\p{Letter}/gu ref: developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/…

[jave.web]

JavaScript needs u after regex to detect the unicode group: /\p{Letter}/gu

[dimitar.bogdanov]

^ or any Cyrillic letters

[Eric Soyke]

For a long time I had been using [A-z]+ but just noticed this allows a few special characters like ` and [ to slip in. [a-zA-Z]+ is indeed the way to go.