In Java, is there a more elegant way to remove duplicate strings from and ArrayList of Strings?

Question

So, a long story short, I have a Java homework assignment that requires a long ArrayList of Strings to be manipulated in various ways (we're doing things like showing combinations of words, adding and removing from the ArrayList, nothing too special). I noticed that a few of the provided ArrayLists have duplicate entries (and the duplicates aren't necessary for this assignment), so I got the okay from my teacher to sanitize the data by removing duplicate entries. Here's what I came up with:

private static ArrayList<String> KillDups(ArrayList<String> ListOfStrings) {  

    for (int i = 0 ; i < ListOfStrings.size(); i++) {
        for (int j = i + 1; j < ListOfStrings.size(); j++) {
            //don't start on the same word or you'll eliminate it.
            if ( ListOfStrings.get(i).toString().equalsIgnoreCase( ListOfStrings.get(j).toString() )  ) {
                ListOfStrings.remove(j);//if they are the same, DITCH ONE.
                j = j -1; //removing the word basically changes the index, so swing down one.
            }                                
        }
    }
    return ListOfStrings;
}

This is fine for my assignment, but I doubt it would be very useful in the real world. Is there a way to do this that would ignore white space and special characters during the comparison? Is there a cleaner way in general to handle this (maybe without the nested For Loops)? Is there another question I should be asking that I don't know to ask?

Answer 1

Yes. And it can be done in just 1 (elegant) line:

List<String> noDups = new ArrayList<String>(new LinkedHashSet<String>(list));

The intermediate Set ensures no duplicates. The LinkedHashSet implementation of Set was chosen to preserve the order of the list.

Also, on a style note:

• name your methods and parameters with names starting with a lowercase letter
• always refer to the abstract (ie List) rather than the concrete (ie ArrayList) when specifying method signatures

Your whole method is then:

private static List<String> killDups(List<String> list) {
    return new ArrayList<String>(new LinkedHashSet<String>(list));
}

For extra brownie points make the method generic, so it works with any type of List:

private static <T> List<T> killDups(List<T> list) {
    return new ArrayList<T>(new LinkedHashSet<T>(list));
}

If you wanted to ignore certain characters, I'd create a class for that and have a list of those. Both the hashCode() and the equals() methods are relied upon by HashSets to remove dups:

public class MungedString {
    // simplified code
    String s;

    public boolean equals(Object o) {
        // implement how you want to compare them here
    }

    public int hashCode() {
        // keep this consistent with equals()
    }
}

then

List<MungedString> list;
List<MungedString> noDupList = killDups(list);

Answer 2

Consider using Set

For the most simple case, which is direct comparison of string, using Hashset is what you would want to do:

Set<String> mySet = new HashSet<String>();

mySet.addAll(aListWithDuplciatedStrings);

then, what's inside mySet will be the unique set of strings.

For ignore-case comparison, it is the homework I left to you. Look at TreeSet and Comparator