Numerals in Unicode

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

Numerals (often called numbers in Unicode) are characters or sequences of characters that denote a number. The Hindu-Arabic numeral system (base-10) is used widely in various writing systems throughout the world and all share the same semantics for denoting numbers. However, the graphemes representing the numerals differ widely from one writing system to another. To support these grapheme differences, Unicode includes encodings of these numerals within many of the script blocks. The decimal digits are repeated in 22 separate blocks. In addition to many forms of the Hindu-Arabic numerals, Unicode also includes several less common numerals such as: Aegean numerals, Roman numerals, counting rod numerals, Cuneiform numerals and ancient Greek numerals. There is also a large number of typographical variations of the Arabic numerals provided for specialized mathematical use and for compatibility with earlier character sets, and also composite characters containing Arabic numerals such as ½.

Numerals invariably involve composition of glyphs as a limited number of characters are composed to make other numerals. For example, the sequence 9–9–0 in Arabic numerals composes the numeral for nine hundred ninety (990). In Roman numerals, the same number is expressed by the composed numeral Ⅹↀ or ⅩⅯ. Each of these is a distinct numeral for representing the same abstract number. The semantics of the numerals differ in particular in their composition. Hindu-Arabic digits are positional-value compositions, while the Roman numerals are sign-value and they are additive and subtractive depending on their composition.

Numerals by numeric property[edit]

Grouped by their numerical property as used in a text, Unicode has four values for Numeric Type. First there is the "not a number" type. Then there are decimal-radix numbers, commonly used in Western style decimals (plain 0-9), there are numbers that are not part of a decimal system such as Roman numbers, and decimal numbers in typographic context, such as encircled numbers. Not noted is a numbering like "A. B. C." for chapter numbering.

Numeric Type[a][b] (Unicode character property)
Numeric type Code Has Numeric Value Example Remarks
Not numeric None No
  • A
  • X (Latin)
  • !
  • Д
  • μ
Numeric Value="NaN"
Decimal De Yes
  • 0
  • 1
  • 9
  •  (Devanagari 6)
  •  (Kannada 6)
  • 𝟨 (Mathematical, styled sans serif)
Straight digit (decimal-radix). Corresponds both ways with General Category=Nd[a]
Digit Di Yes
  • ¹ (superscript)
  •  (digit with full stop)
Decimal, but in typographic context
Numeric Nu Yes
  • ¾
  •  (Tamil number ten)
  •  (Roman numeral)
  •  (Han number 6)
Numeric value, but not decimal-radix
a. ^ "Section 4.6: Numeric Value" (PDF). The Unicode Standard. Unicode Consortium. July 2018.
b. ^ "Unicode 11.0 Derived Numeric Types". Unicode Character Database. Unicode Consortium. 2018-01-31.

Hexadecimal digits[edit]

Hexadecimal digits in Unicode are not separate characters, existing letters and numbers are used. These characters have marked Character properties Hex_digit=Yes, and ASCII_Hex_digit=Yes when appropriate.

Characters in Unicode marked Hex_Digit=Yes[a]
0123456789ABCDEF Basic Latin, capitals Also ASCII_Hex_Digit=Yes
0123456789abcdef Basic Latin, small letters Also ASCII_Hex_Digit=Yes
0123456789ABCDEF Fullwidth forms, capitals
0123456789abcdef Fullwidth forms, small letters
a. ^ "Unicode 11.0 UCD: PropList.txt". 2018-03-15. Retrieved 2018-06-05.

Numerals by script[edit]

Hindu-Arabic numerals[edit]

The Hindu-Arabic numeral system involves ten digits representing 0–9. Unicode includes the Western Arabic numerals in the Basic Latin (or ASCII derived) block. The digits are repeated in several other scripts: Eastern Arabic, Balinese, Bengali, Devanagari, Ethiopic, Gujarati, Gurmukhi, Telugu, Khmer, Lao, Limbu, Malayalam, Mongolian, Myanmar, New Tai Lue, Nko, Oriya, Telugu, Thai, Tibetan, Osmanya. Unicode includes a numeric value property for each digit to assist in collation and other text processing operations. However, there is no mapping between the various related digits.

Note that although Arabic is written from right to left, while English is written left to right, in both languages numbers are written with the most significant digit on the left and the least significant on the right.


The fraction slash character (U+2044) allows authors using Unicode to compose any arbitrary fraction along with the decimal digits. This was intended to instruct font rendering to make the surrounding digits smaller and raise them on the left and lower them on the right, but this is rarely implemented (a workaround is to use the super/subscript characters described below but only Arabic numerals are available). Unicode also includes a handful of vulgar fractions as compatibility characters, but discourages their use.

Decimal fractions[edit]

Several characters in Unicode can serve as a decimal separator depending on the locale. Decimal fractions are represented in text as a sequence of decimal digit numerals with a decimal separator separating the whole-number portion from the fractional portion. For example, the decimal fraction for “¼” is expressed as zero-point-two-five (“0.25”). Unicode has no dedicated general decimal separator but unifies the decimal separator function with other punctuation characters. So the “.” used in “0.25” is the same period character (U+002E) used to end the sentence. However, cultures vary in the glyph or grapheme used for a decimal separator. So in some locales, the comma (U+002C) may be used instead ”0,25”. Still other locales use a space (or non-breaking space) for “0 25”. The Arabic writing system includes a dedicated character for a decimal separator that looks much like a comma ”٫” (U+066B) which when combined with the Arabic digits to express one-quarter appears as: “٠٫٢٥”.

Characters for mathematical constants[edit]

Currently, three Unicode characters semantically represent mathematical constants: U+210E PLANCK CONSTANT, the U+210F PLANCK CONSTANT OVER TWO PI, and U+2107 EULER CONSTANT. Other mathematical constants can be represented using characters that have multiple semantic uses. For example, although Unicode includes a character for natural exponent ℯ (U+212F) its UCS canonical name derives from its glyph: U+212F SCRIPT SMALL E; and the mathematical constant π, 3.141592.., is represented by U+03C0 π GREEK SMALL LETTER PI.

Rich text and other compatibility numerals[edit]

The Western Arabic numerals also appear among the compatibility characters as rich text variant forms including bold, double-struck, monospace, sans-serif and sans-serif bold. and fullwidth variants for legacy vertical text support.

Rich text parenthesized, circled and other variants are also included in the blocks: Enclosed CJK Letters and Months; Enclosed Alphanumerics, Superscripts and Subscripts; Number Forms; and Dingbats.

CJK Suzhou (huāmǎ) numerals[edit]

The huāmǎ system is a variation of the rod numeral system. Rod numerals are closely related to the counting rods and the abacus, which is why the numeric symbols for 1, 2, 3, 6, 7 and 8 in the huāmǎ system are represented in a similar way as on the abacus. Nowadays, the huāmǎ system is only used for displaying prices in Chinese markets or on traditional handwritten invoices.

Suzhou (huāmǎ) numerals in Unicode[edit]

According to the Unicode standard version 3.0, these characters are called Hangzhou style numerals. This indicates that it is not used only by Cantonese in Hong Kong. In the Unicode standard 4.0, an erratum was added which stated:

The digits of the Suzhou numerals are designated in the CJK Symbols and Punctuation block between U+3021 and U+3029, U+3007, U+5341, U+5344, and. U+5345.

Japanese and Korean numerals[edit]

Ancient Greek numerals[edit]

Unicode provides support for several variants of Greek numerals, assigned to the Supplementary Multilingual Plane from U+10140 through U+1018F.[1]

Attic numerals were used by ancient Greeks, possibly from the 7th century BC. They were also known as Herodianic numerals because they were first described in a 2nd-century manuscript by Herodian. They are also known as acrophonic numerals because all of the symbols used derive from the first letters of the words that the symbols represent: 'one', 'five', 'ten', 'hundred', 'thousand' and 'ten thousand'. See Greek numerals and acrophony.

Decimal Symbol Greek numeral
1 Ι ἴος or ἰός (ios)
5 Π πέντε (pente)
10 Δ δέκα (deka)
100 Η ἑκατόν (hekaton)
1000 Χ χίλιοι (khilioi)
10000 Μ μύριοι (myrioi)
Ancient Greek Numbers[1][2]
Official Unicode Consortium code chart (PDF)
  0 1 2 3 4 5 6 7 8 9 A B C D E F
U+1014x 𐅀 𐅁 𐅂 𐅃 𐅄 𐅅 𐅆 𐅇 𐅈 𐅉 𐅊 𐅋 𐅌 𐅍 𐅎 𐅏
U+1015x 𐅐 𐅑 𐅒 𐅓 𐅔 𐅕 𐅖 𐅗 𐅘 𐅙 𐅚 𐅛 𐅜 𐅝 𐅞 𐅟
U+1016x 𐅠 𐅡 𐅢 𐅣 𐅤 𐅥 𐅦 𐅧 𐅨 𐅩 𐅪 𐅫 𐅬 𐅭 𐅮 𐅯
U+1017x 𐅰 𐅱 𐅲 𐅳 𐅴 𐅵 𐅶 𐅷 𐅸 𐅹 𐅺 𐅻 𐅼 𐅽 𐅾 𐅿
U+1018x 𐆀 𐆁 𐆂 𐆃 𐆄 𐆅 𐆆 𐆇 𐆈 𐆉 𐆊 𐆋 𐆌 𐆍 𐆎
1.^ As of Unicode version 11.0
2.^ Grey area indicates non-assigned code point

Roman numerals[edit]

Roman numerals are a numeral system originating in ancient Rome, adapted from Etruscan numerals. The system used in classical antiquity was slightly modified in the Middle Ages to produce the system we use today. It is based on certain letters which are given values as numerals.

Roman numerals are commonly used today in numbered lists (in outline format), clockfaces, pages preceding the main body of a book, chord triads in music analysis (Roman numeral analysis), the numbering of movie and video game sequels, book publication dates, successive political leaders or children with identical names, and the numbering of some sport events, such as the Olympic Games or the Super Bowl.

Roman numerals in Unicode[edit]

Unicode has a number of characters specifically designated as Roman numerals, as part of the Number Forms[2] range from U+2160 to U+2188. This range includes both upper- and lowercase numerals, as well as pre-combined characters for numbers up to 12 (Ⅻ or XII). One reason for the existence of pre-combined numbers is to facilitate the setting of multiple-letter numbers (such as VIII) on a single horizontal line in Asian vertical text. The Unicode standard, however, includes special Roman numeral code points for compatibility only, stating that "[f]or most purposes, it is preferable to compose the Roman numerals from sequences of the appropriate Latin letters".[3]

Additionally, characters exist for archaic[2] forms of 1000, 5000, 10,000, large reversed C (Ɔ), late 6 (ↅ, similar to Greek Stigma: Ϛ), early 50 (ↆ, similar to down arrow ↓⫝⊥[4]), 50,000, and 100,000. Note that the small reversed c, ↄ is not intended to be used in Roman numerals, but as lower case Claudian letter Ↄ,

Table of Roman numerals in Unicode
Code x= 0 1 2 3 4 5 6 7 8 9 A B C D E F
Value[5] 1 2 3 4 5 6 7 8 9 10 11 12 50 100 500 1,000
Value 1000 5000 10,000 6 50 50,000 100,000

If using blackletter or script typefaces, Roman numerals are set in Roman type. Such typefaces may contain Roman numerals matching the style of the typeface in the Unicode range U+2160–217F; if they don't exist, a matching Antiqua typeface is used for Roman numerals.

Unicode has characters for Roman fractions in the Ancient Symbols[6] block: sextans, uncia, semuncia, sextula, dimidia sextula, siliqua, and as.

Counting-rod numerals[edit]

Value 0 1 2 3 4 5 6 7 8 9
Vertical Counting rod 0.png Counting rod v1.png Counting rod v2.png Counting rod v3.png Counting rod v4.png Counting rod v5.png Counting rod v6.png Counting rod v7.png Counting rod v8.png Counting rod v9.png
Horizontal Counting rod 0.png Counting rod h1.png Counting rod h2.png Counting rod h3.png Counting rod h4.png Counting rod h5.png Counting rod h6.png Counting rod h7 num.png Counting rod h8 num.png Counting rod h9 num.png

The vertical rods are usually for even powers of ten (1, 100, 10000...) and the horizontal for odd powers (10, 1000...). For example, 126 is represented by Counting rod v1.pngCounting rod h2.pngCounting rod v6.png instead of Counting rod v1.png Counting rod v2.pngCounting rod v6.png, which could be confused with 36. Historically, red rods were used for positive numbers and black rods for negative numbers.

Counting rod numerals in Unicode[edit]

Counting rod numerals are included in their own block in the Supplementary Multilingual Plane (SMP) from U+1D360 to U+1D37F. Eighteen characters for vertical and horizontal digits of 1-9 are included as of Unicode 5.0, though vertical and horizontal are opposite from the description above. Fourteen code points reserved for future use. Zero should be represented by U+3007 (〇, ideographic number zero) and the negative sign should be represented by U+20E5 (combining reverse solidus overlay).[7] As these were recently added to the character set and since they are included in the SMP, font support may still be limited.

Counting Rod Numerals[1][2]
Official Unicode Consortium code chart (PDF)
  0 1 2 3 4 5 6 7 8 9 A B C D E F
U+1D36x 𝍠 𝍡 𝍢 𝍣 𝍤 𝍥 𝍦 𝍧 𝍨 𝍩 𝍪 𝍫 𝍬 𝍭 𝍮 𝍯
U+1D37x 𝍰 𝍱 𝍲 𝍳 𝍴 𝍵 𝍶 𝍷 𝍸
1.^ As of Unicode version 11.0
2.^ Grey areas indicate non-assigned code points

See also[edit]