Add number formatting support for locales that use non-uniform digit grouping (e.g. India)

[Binoj David]

When using 'en-IN' as locale (listed as valid in ​http://www.i18nguy.com/unicode/language-identifiers.html) 100000.00 should look like 1,00,000.00 after using intcomma. Instead 100,000.00 is displayed.

Works as expected in locale package as below (with en_IN installed, of course):

In [1]: import locale

In [2]: locale.setlocale(locale.LC_ALL, 'en_IN')
Out[2]: 'en_IN'

In [3]: locale.format("%.2f", 100000.00, grouping=True)
Out[3]: '1,00,000.00'

[Tim Graham]

I think the solution here is to add formats for the en_IN locale (like ff2e0896a3f7b566c8a0802b2191d30c6070c8f6, for example). Please send a pull request if you can do that. I don't think there's much value in leaving an open ticket for each locale that we're missing. Thanks!

[Binoj David]

There is a complication in adding en_IN. Number grouping in en_IN is not grouped on 'n' number of digits. First 3 (ones, tens and hundreds) are grouped as a single group and the remaining digits are grouped as 2 digits per group. This can be achieved with a NUMBER_GROUPING setting. So I don't think this can be fixed by adding a locale.

[Claude Paroz]

Confirmed, Django doesn't support this sort of grouping yet.

[Szymon Teżewski]

PR for this: ​https://github.com/django/django/pull/6693
It basically adds possibility to use more robust NUMBER_GROUPING using style similar to POSIX locale grouping option.

[Philip James]

I looked at the pull request, and think it looks RTC for the following reasons:

• Has tests
• Test suite passes
• Old functionality from previous tests remains unchanged.

[Tim Graham]

A bit of documentation is missing as noted on the PR.

[Tim Graham <timograham@…>]

In b5a1c3a:

Fixed #25920 -- Added support for non-uniform NUMBER_GROUPING.