Unicode Categories — \p{...}

Match by Unicode property, not by listing characters

\p{...} matches any character with the given Unicode property. The complement is \P{...}. This unlocks portable Unicode-aware patterns without listing every character.

Common properties

\p{L} — any letter (any script)
\p{Ll} — lowercase letter
\p{Lu} — uppercase letter
\p{N} — any number-like (digits, fractions, Roman numerals, etc.)
\p{Nd} — decimal digit (any script's digits)
\p{P} — punctuation
\p{S} — symbol (math, currency, modifier)
\p{Z} — separator (space, line)
\p{Mn} — non-spacing mark (combining diacritics)

Script properties

\p{Script=Hangul} matches Korean Hangul syllables. \p{Script=Han} matches CJK ideographs. \p{Script=Latin} for Latin script. Useful for filtering by writing system.

Engine support

Python regex third-party: Full support.
Python re: No \p{...}, but \d, \w, \s are Unicode-aware by default.
JavaScript: Yes, with /u flag.
PCRE, Java, .NET, Ruby: Yes.
Go RE2: Yes (uses RE2 syntax).

When to use

Anytime you're processing multilingual text. [\p{L}\p{N}_] is a portable Unicode-aware identifier pattern, working for Korean variable names, Cyrillic, Arabic, anything. [a-zA-Z0-9_] only works for ASCII.

Code

Unicode property patterns·python

# Python — needs the third-party 'regex' module for \p{...}
import regex

# All letters in any script
regex.findall(r'\p{L}+', 'hello 안녕 123')
# ['hello', '안녕']

# Korean Hangul only
regex.findall(r'\p{Script=Hangul}+', 'hello 안녕 こんにちは')
# ['안녕']

# CJK ideographs
regex.findall(r'\p{Script=Han}+', '中文 한자 漢字')
# ['中文', '漢字']  — '한자' is Hangul phonetic, not Han

# Identifier pattern (Unicode-aware)
regex.findall(r'[\p{L}\p{N}_]+', 'var_name = 변수명 + 한글_var')
# ['var_name', '변수명', '한글_var']

# Strip combining marks (e.g., remove diacritics from text)
import unicodedata
clean = regex.sub(r'\p{Mn}', '',
                  unicodedata.normalize('NFD', 'café résumé'))
# 'cafe resume'

# JavaScript equivalent (with /u)
// '안녕 hello'.match(/\p{Script=Hangul}+/gu)
// ['안녕']

Unicode Categories — \p{...}

Match by Unicode property, not by listing characters

Common properties

Script properties

Engine support

When to use

Code

External links

Exercise

Progress

Comments 0