Thomas Wolff <[email protected]> writes: > Am 07.04.2026 um 12:28 schrieb Dan Jacobson: >> I hereby propose Coreutils' sort(1) add the ability to sort Chinese >> (actually CJK) numbers. >> >> https://chinese.stackexchange.com/questions/64035/how-to-sort-chinese-numbers-with-a-computer >> >> Isn't Chinese has the most native speakers in the world so it's high >> time that sort(1) deal with the numbers, pun intended. > A suitable basis for such handling is file Unihan_NumericValues.txt in > the Unihan.zip from Unicode.org.
GNU libunistring has the uc_numeric_value function to convert Unicode characters to numeric values. I had previously considered proposing the functionality to 'numfmt'. I'm not sure it is worth adding to 'sort'. My guess is that it will not be used very frequently, but perhaps I am wrong. If the feature were added, though, there would certainly be no point in limiting things to Chinese numerals. There are many other symbols used worldwide, see the many used in India alone for example [1]. Collin [1] https://en.wikipedia.org/wiki/Hindu%E2%80%93Arabic_numeral_system#Symbols
