Numbers

import {
  convertNumbers,
  toUrduDigits,
  toEnglishDigits,
  toArabicIndicDigits,
  parseUrduNumber,
  numberToUrduWords,
} from "urdu-text-utils";

Two digit systems, not one

System	Digits	Block
Urdu / extended Arabic-Indic	۰۱۲۳۴۵۶۷۸۹	U+06F0
Arabic-Indic	٠١٢٣٤٥٦٧٨٩	U+0660
ASCII	0123456789	U+0030

Urdu content uses the first. Content copied from Arabic sources uses the second. They render similarly and compare as completely different characters, so a phone number or an invoice total can arrive in either.

Every function here accepts all three as input.

Converting

convertNumbers("12345");            // "۱۲۳۴۵"
convertNumbers("۱۲۳۴۵", "english"); // "12345"

toUrduDigits("١٢٣");        // "۱۲۳" — Arabic-Indic in, Urdu out
toEnglishDigits("۳۱-۱۲-۲۰۲۴"); // "31-12-2024"
toArabicIndicDigits("123"); // "١٢٣"

Non-digits pass through untouched, so it is safe on whole sentences:

toUrduDigits("سال 2024ء میں"); // "سال ۲۰۲۴ء میں"

Prefer the named functions over convertNumbers(text, style) in new code — they are typed without a magic string and tree-shake individually.

Parsing

Number("۱۲۳") is NaN. Use:

parseUrduNumber("۱۲۳۴۵"); // 12345
parseUrduNumber("۱٬۲۳۴");  // 1234 — ٬ is the thousands separator
parseUrduNumber("۳٫۱۴");   // 3.14 — ٫ is the decimal separator
parseUrduNumber("پاکستان"); // NaN

Note ٫ (U+066B) and ٬ (U+066C) are not the ASCII . and ,. Both are handled, along with ASCII separators and surrounding whitespace.

Numbers in words

numberToUrduWords(5);        // "پانچ"
numberToUrduWords(100);      // "ایک سو"
numberToUrduWords(100000);   // "ایک لاکھ"
numberToUrduWords(10000000); // "ایک کروڑ"
numberToUrduWords(-5);       // "منفی پانچ"

It uses the South Asian scale — ہزار, لاکھ, کروڑ, ارب — rather than thousand/million/billion, which is what Urdu readers expect for prices and populations.

Experimental

Compound numbers between 21 and 99 are irregular in Urdu (اکیس, بیالیس, چھیاسٹھ are not composed from their parts). Multiples of ten and the teens are correct; other values fall back to a composed form that reads awkwardly. Fine for approximate labels, not for legal or financial text.

Try it

normalizeUrdu(text) stripDiacritics

Input

كيا حال ہے

Output

کیا حال ہے