ค่าตัวเลข

What Is the Numeric Value Property?

Unicode assigns numeric values to characters that represent numbers across many writing systems. The Numeric_Value property stores the actual number (as a rational), and the companion Numeric_Type property classifies how the character represents a number. This covers far more than the ASCII digits 0–9: Arabic-Indic digits, Roman numerals, circled numbers, vulgar fractions, and Counting Rod numerals all carry numeric values.

Numeric_Type Values

import unicodedata

chars = [
    ("5",      "ASCII five"),
    ("٥",      "Arabic-Indic five"),
    ("²",      "Superscript two"),
    ("½",      "Vulgar fraction one half"),
    ("Ⅻ",      "Roman numeral 12"),
    ("〇",     "CJK zero"),
    ("A",      "Letter A (not numeric)"),
]

for char, label in chars:
    decimal = unicodedata.decimal(char, None)
    digit   = unicodedata.digit(char, None)
    numeric = unicodedata.numeric(char, None)
    print(f"  {char}  {label}")
    print(f"    decimal={decimal}  digit={digit}  numeric={numeric}")

# 5  decimal=5  digit=5  numeric=5.0
# ٥  decimal=5  digit=5  numeric=5.0
# ²  decimal=None  digit=2  numeric=2.0
# ½  decimal=None  digit=None  numeric=0.5
# Ⅻ decimal=None  digit=None  numeric=12.0
# 〇 decimal=None  digit=None  numeric=0.0
# A  decimal=None  digit=None  numeric=None

The Fraction Challenge

Vulgar fractions like U+2153 (⅓, ONE THIRD) have Numeric_Value = 1/3. Python's unicodedata.numeric() returns the Python float approximation (0.3333...). For exact rational arithmetic you need the stored rational value from the Unicode Character Database directly. The Decimal type is not used here; the float precision is usually sufficient for display but not for precise computation.

Practical Uses

• Form validation: allow digit characters from any script in numeric fields by checking unicodedata.decimal(c, None) is not None.
• Sorting numbers expressed in text: normalise digits to ASCII before comparison.
• OCR post-processing: map non-ASCII digits to their ASCII equivalent using unicodedata.decimal().
• Accessibility: screen readers use Numeric_Value to pronounce fraction characters correctly.

Quick Facts