How to get the ordinal suffix of a number for many languages in Java or Groovy

Question

I am building a multi-languages grails website and I need to get the ordinal suffix of a number for several languages like English, French, Spanish, German and Italian.

I believe that this issue is quite common for multi-languages website owners. I have found this article that provides a solution but it is English only.

For instance:

/**
 *@value number 
 *@locale Current locale
 *@returns ordinal suffix for the given number
**/
public static String getOrdinalFor(int value, Locale locale) 

will give the following results:

 assert getOrdinalFor(1, Locale.ENGLISH) == "st"
 assert getOrdinalFor(1, Locale.FRENCH) == "er"
 assert getOrdinalFor(2, Locale.ENGLISH) == "nd"
 assert getOrdinalFor(3, Locale.ENGLISH) == "rd"
 assert getOrdinalFor(4, Locale.ENGLISH) == "th"
 assert getOrdinalFor(4, Locale.FRENCH) == "ème"

Do you know a library (in Java or Groovy) that can help this? Or do you know an algorithm that implements it?

Answer 1

I think in many languages this approach is impossible as they simply have no concept of ordinals that can be composed of a number with some letters as extension. For "1st" i.e. in german you can only either write "1." or "erster"/"erste"/"erstes" depending of genus of what you are numbering.

Answer 2

I don't know of any library, but the best algorithm may be to get hold of all possible numbers/locales/ordinals in a Map<Integer, Map<Locale, String>>.

private static Map<Integer, Map<Locale, String>> ordinals = new HashMap<Integer, Map<Locale, String>>();
static {
    Map<Locale, String> ordinal1 = new HashMap<Locale, String>();
    ordinal1.put(Locale.ENGLISH, "st");
    ordinal1.put(Locale.FRENCH, "er");
    ordinals.put(1, ordinal1);
    Map<Locale, String> ordinal2 = new HashMap<Locale, String>();
    ordinal2.put(Locale.ENGLISH, "nd");
    ordinals.put(2, ordinal2);
    Map<Locale, String> ordinal3 = new HashMap<Locale, String>();
    ordinal3.put(Locale.ENGLISH, "rd");
    ordinals.put(3, ordinal3);
    Map<Locale, String> ordinal4 = new HashMap<Locale, String>();
    ordinal4.put(Locale.ENGLISH, "th");
    ordinal4.put(Locale.FRENCH, "ème");
    ordinals.put(4, ordinal4);
}

The getOrdinalFor() could then basically look like (obvious runtime error handling aside):

public static String getOrdinalFor(int value, Locale locale) {
    return ordinals.get(value).get(locale);
}

For other numbers which are not covered by the map, you may consider to countdown the number until a match is found in the map (so that requesting the ordinal for e.g. 5 or higher would return the ordinal for 4 (the firstnext match on countdown)):

public static String getOrdinalFor(int value, Locale locale) {
    Map<Locale, String> ordinal = null;
    for (int i = value; i > 0 && (ordinal == null || !ordinal.containsKey(locale)); i--) {
        ordinal = ordinals.get(i);
    }
    return ordinal != null ? ordinal.get(locale) : null; // Or other default? 
}

Answer 3

Unless you can apply a rule across every language, it is common to try to sidestep complex linguistic issues. You have to consider not only the number and its ordinal, but the context in which it appears and even how position within a sentence might affect translation.

Rather than try to write "1st place", a resource like this is much easier to translate:

#easyToTranslate_en.properties
label.position=Position: {0,number,integer}

Depending on what you're trying to do, you may be able to compromise with a solution like this:

#compromise_en.properties
label.position.first=1st place
label.position.second=2nd place
label.position.third=3rd place
label.position.default=Position: {0,number,integer}

This would require the code reading the resource to perform a case test to choose which resource to load.