标点符号

用于组织和阐明书面语言的字符，包括句号、逗号、破折号、引号等，Unicode一般类别P涵盖所有标点符号。

2022-07-30 · 更新日期 2024-12-19

What Is Unicode Punctuation?

Unicode organizes punctuation characters under the top-level General Category P (Punctuation), subdivided into six specific categories. Punctuation encompasses marks used to structure written language: paired delimiters, dashes, connectors, and miscellaneous marks that vary widely across the world's writing systems.

Unlike Latin-centric definitions of punctuation, Unicode's coverage includes CJK ideographic punctuation, Arabic quotation marks, Ethiopic wordspace, and hundreds of other culturally specific marks.

The Six Punctuation Subcategories

Code	Name	Examples
Pc	Connector Punctuation	_ (LOW LINE), ‿ (UNDERTIE)
Pd	Dash Punctuation	- (HYPHEN-MINUS), – (EN DASH), — (EM DASH), ― (HORIZONTAL BAR)
Ps	Open Punctuation	( [ { ⟨ 「《
Pe	Close Punctuation	) ] } ⟩ 」》
Pi	Initial Punctuation	" (LEFT DOUBLE QUOTATION MARK) « (LEFT-POINTING DOUBLE ANGLE QUOTATION MARK)
Pf	Final Punctuation	" (RIGHT DOUBLE QUOTATION MARK) » (RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK)
Po	Other Punctuation	! . , : ; ? @ # % & * / \ (and many more)

import unicodedata

punctuation_samples = [
    ("_",  "Pc - connector"),
    ("-",  "Pd - hyphen-minus"),
    ("–",  "Pd - en dash U+2013"),
    ("(",  "Ps - open paren"),
    (")",  "Pe - close paren"),
    ("「", "Ps - CJK left corner bracket"),
    ("」", "Pe - CJK right corner bracket"),
    ("\u201C", "Pi - left double quotation"),
    ("\u201D", "Pf - right double quotation"),
    ("«",  "Pi - left-pointing angle quotation"),
    ("»",  "Pf - right-pointing angle quotation"),
    ("。", "Po - CJK ideographic full stop"),
    ("，", "Po - fullwidth comma"),
    ("!",  "Po - exclamation mark"),
]

for char, label in punctuation_samples:
    gc = unicodedata.category(char)
    print(f"  U+{ord(char):04X}  {char}  {gc}  {label}")

CJK and Fullwidth Punctuation

East Asian writing uses a distinct set of punctuation marks designed for fullwidth (double-byte) layout:

。 (U+3002) IDEOGRAPHIC FULL STOP — sentence terminator in Chinese and Japanese
、 (U+3001) IDEOGRAPHIC COMMA — list separator
「」 (U+300C, U+300D) — Japanese corner brackets for quotation
『』 (U+300E, U+300F) — double corner brackets for nested quotation
· (U+00B7) vs ・ (U+30FB) — Latin middle dot vs Katakana middle dot

Fullwidth versions of ASCII punctuation (U+FF01–U+FF0F) are compatibility equivalents of their halfwidth counterparts; NFKC normalization maps them back to ASCII.

Distinguishing Similar Punctuation

Many punctuation characters look similar but have different code points, names, and semantic roles:

Character	Code Point	Name
-	U+002D	HYPHEN-MINUS
‐	U+2010	HYPHEN
–	U+2013	EN DASH
—	U+2014	EM DASH
―	U+2015	HORIZONTAL BAR
'	U+0027	APOSTROPHE
'	U+2018	LEFT SINGLE QUOTATION MARK
'	U+2019	RIGHT SINGLE QUOTATION MARK

Smart quote substitution in word processors converts ASCII ' and " to their Pi/Pf counterparts. This can cause problems in code contexts where the apostrophe is syntactically significant.

Quick Facts

Property	Value
General Category group	P (Punctuation)
Subcategories	Pc, Pd, Ps, Pe, Pi, Pf, Po (7)
Python function	`unicodedata.category(char)` starting with `P`
Regex	`\p{Punctuation}` or `\p{P}` (PCRE/Python regex)
CJK punctuation block	U+3000–U+303F (CJK Symbols and Punctuation)
Fullwidth punctuation	U+FF01–U+FF60 (compatibility equivalents)
Spec reference	Unicode Standard Chapter 6