Tjyylli

I have following Python code:

def get_subject_from_stream_id_and_subject_id(stream_id, subject_id):

    #(stream_id, subject_id): ("subject_name")

    return {

        (1, 1): "Accounts",

        (1, 2): "English",

        (1, 3): "Organization of Commerce",

        (2, 1): "Physics",

        (2, 2): "English",

        (2, 3): "Biology"

    }.get((stream_id, subject_id), "None")

In this code, I want to get subject name from the integer pair combination i.e. stream_id, subject_id e.g. (1, 2) is for English. It was implemented using a Python tuple.

I want to implement the same piece of code in Java.

Could someone write this in a better way in Java?

public String getSubjectFromStreamIdAndSubjectId(int streamId, int subjectId) {

  switch (streamId) {

    case 1:

        switch (subjectId) {

        case 1:

            return "Accounts";

        case 2:

            return "English";

        case 3:

            return "Organization of Commerce";

        default:

            return null;

        }



    case 2:

        switch (subjectId) {

        case 1:

            return "Physics";

        case 2:

            return "English";

        case 3:

            return "Biology";

        default:

            return null;

        }

    default:

        return null;

   }

}

Java equivalent of the above Python code goes more like this:

private static final Map<List<Integer>, String> SUBJECT_MAP = createSubjectMap();



private static Map<List<Integer>, String> createSubjectMap() {

    Map<List<Integer>, String> map = new HashMap<>();

    map.put(asList(1, 1), "Accounts");

    map.put(asList(1, 2), "English");

    map.put(asList(1, 3), "Organization of Commerce");

    map.put(asList(2, 1), "Physics");

    map.put(asList(2, 2), "English");

    map.put(asList(2, 3), "Biology");

    return map;

}



public static String getSubjectFromStreamIdAndSubjectId(int streamId, int subjectId) {

    return SUBJECT_MAP.get(asList(streamId, subjectId));

}

Neither is good, but surely we can all agree that the Java example is way uglier. The biggest problem, however, is that both examples hard code data into code.

Separate data, i.e. stream and subject IDs and their titles into a data class and in that class implement code that accesses the data structure without detailed knowledge about the actual data. Responsibility of setting up the data structure to resemble your stream and subject numbering is left to a separate component (load it from file or set up in static code).

Whether the data class is a recursive structure or just a wrapper for a HashMap depends on the complexity of your data.

When something is difficult in Java it's nearly always because there is a better way to do it (in Java, it generally means you are missing a class). In your case, it looks like you need a "Subject" class.

Let's say you had a "Subject" class, how would it be implemented? Here's one way (Not my favorite)

enum Subject {

    Accounts(1,1),

    English(1,2),

    …;

   public static Subject getSubject(streamId, subjectId) {

       // Iterate over Subject.values and return one that matches

   }

This means that your call becomes simple:

assertEquals(Subjects.Accounts, Subject.getSubject(1,1));

I don't completely love this because it's still hard-coded. I would personally make the Subjects class a full class and load it from either a database or a data file, but it should still have a getSubject() static.

Just in case you think there is some kind of performance issue caused by looping instead of a switch, note that A) java is 10x faster than python to start and B) premature optimization is the root of all evil (well, lots of evil)