数字字符引用

What Are Numeric Character References?

Numeric character references (NCRs) are HTML escape sequences that represent any Unicode character by its code point number. They take two forms:

• Decimal: &#N; where N is a base-10 integer — e.g., © for ©
• Hexadecimal: &#xH; where H is a base-16 integer — e.g., © for ©

Both forms refer to the Unicode scalar value of the character. Since Unicode covers over 1.1 million code points, NCRs can represent virtually any character ever assigned — from basic Latin letters to rare CJK ideographs and emoji — using only ASCII characters in the source.

Decimal vs. Hexadecimal

Decimal NCRs (&#N;) are straightforward for readers who know the decimal code points of common characters (65 = 'A', 169 = '©'). Hexadecimal NCRs (&#xH;) align with how Unicode code points are conventionally written — U+00A9 maps directly to ©. When working with Unicode documentation or character tables that list code points in hex, the hex form is easier to use without mental conversion.

<!-- These are identical -->
&#65;     = &#x41;   = A
&#169;    = &#xA9;   = ©
&#8364;   = &#x20AC; = €
&#128512; = &#x1F600; = 😀

Valid Range

Valid code points for NCRs are: 1–55295 (U+0001–U+D7FF) and 57344–1114111 (U+E000–U+10FFFF). The surrogate range U+D800–U+DFFF is invalid and must not be encoded. U+0000 (NULL) is also excluded. Browsers may render other disallowed code points (such as U+0001–U+001F control characters) as the replacement character U+FFFD.

Supplementary Characters

NCRs fully support Unicode supplementary characters (code points above U+FFFF). In UTF-16 these require surrogate pairs, but in HTML you write a single NCR:

<!-- U+1F4A9 PILE OF POO — supplementary character -->
&#128169;      <!-- decimal -->
&#x1F4A9;      <!-- hex -->

This is one advantage of NCRs over raw UTF-16 encoding in old environments.

Practical Use

<!-- Escaping in content -->
<p>The formula is a &#60; b &lt; c</p>
<!-- &#60; and &lt; both render as < -->

<!-- Characters outside keyboard reach -->
<p>The currency symbol is &#x20B9; (Indian Rupee)</p>

<!-- In HTML attributes -->
<input placeholder="Enter &#x2764; here">

<!-- Emoji -->
<title>Unicode Guide &#x1F4DA;</title>

# Python: convert character to NCR
char = "©"
f"&#{ord(char)};"    # "©"
f"&#x{ord(char):X};" # "©"

# Python: decode NCR
import html
html.unescape("©")   # "©"
html.unescape("©")   # "©"

NCRs vs. Named References vs. Direct Characters

In modern UTF-8 documents, direct characters are preferred. NCRs remain valuable in legacy ASCII environments and when generating HTML programmatically.

Quick Facts