Wanted to know if suffixing ordinal indicators is a good idea in applications that are designed to cater to a multi-lingual audience. If we come up with an angular pipe with logic which transforms numbers like 1 -> 1st, 2-> 2nd, 3 -> 3rd rest all suffixed with 'th', how good is this assumption for other languages if we read 'st','nd','rd','th' from a file suffixed with locale identifier like 'en','ar' etc..
If I understand correctly, you're asking if a simple "translating" of suffixes will work. e.g. "st" (in English) to "er" (for premier in French). If that's what you mean then no, it's no good. The rules are too complex and vary too much between languages.
If you only need this function for dates then you have the luxury of only needing 31 pre-defined values. Then you could pre-compile some lists, like these:
en:
["","1st","2nd","3rd", ... ,"29th","30th","31st"]
fr:
["","1er","2e","3e", ..., "29e","30e","31e"]
ar:
["","١.","٢.","٣." , .. ,"٢٩.","٣٠.","٣١."]
I generated these with a simple PHP script using NumberFormatter. I can't vouch for the Arabic being correct.
If you need these suffixes for any arbitrary number, then you'd have to embed the rules into your application.
Ordinal rules are defined in the Unicode CLDR. https://unicode.org/cldr/charts/latest/supplemental/language_plural_rules.html
You could extract these rules from the data files and bake the equations into your application, but this might be a lot of work.
They are also imperfect. I don't remember much French from school but a quick Google shows that (for 1st) you have 1er (premier), 1re (première), 1ers (premiers) and 1res (premières). What a nightmare.
The best route for you would depend on the precise use case, which your question doesn't give.