Comparing elements in a Collection

[Songezo Nkukwana]

Might be a beginner's question ... I'm trying to remove duplicates in a list but don;t know how to compare a list...

Here' my attempt

protected List<NonQualMembersDTO> removeDuplicates(List<NonQualMembersDTO> members) 
{

for(NonQualMembersDTO dto : members) 
{
String surname = dto.getSurname();
Iterator<NonQualMembersDTO> iter = members.iterator();
iter.next();

/* Do we have the same person as next? */
if(surname.equals(iter.next().getSurname())) {

/** REMOVE TO BE INSERTED HERE ... */
}
}

return members;
}

Members is the List which contains people sorted by surnames. This algorithm could be illogical, is there a way i could do this easier?

[Campbell Ritchie]

Why are you returning a List in the first place? To the Java Tutorials, where you find there are lots of different kinds of Collection, some which permit duplicates and some which don't. Using a different kind of Collection might sort out your problem.

But this depends on the equals() and hashCode() methods being correctly overridden in your class.

[Songezo Nkukwana]

Hi Campbell i tried this ..

protected List<NonQualMembersDTO> removeDuplicates(List<NonQualMembersDTO> members) 
{
Set<NonQualMembersDTO> replica = new LinkedHashSet<NonQualMembersDTO>();

for(NonQualMembersDTO dto : members) {
replica.add(dto);
}
 
List<NonQualMembersDTO> newList = new ArrayList<NonQualMembersDTO>();
newList.addAll(replica);
return newList;
}

With the hope that the Set will not permit dupicates, but i still get the same results

[Campbell Ritchie]

That looks correct to me. Have a look at the API for LinkedHashSet and find the add method, and that gives the details. Do you really need to return a List; can you return a Set?

Change the body of the for-each loop to read this, temporarily, as a testing strategy:

System.out.printf("%b ", replica.add(dto));//test

That ought to print true whenever you add, and false whenever you encounter a duplicate. If it won't print false anywhere, check the equals() and hashCode() methods of the DTO class to verify that they have been overridden correctly. They must fulfil the specification of the Object class.

[Peter Lawrey]

Or try

public static <T> List<T> unique(List<T> elements) {
  return new ArrayList<T>(new LinkedHashSet<T>(elements));
}

This assumes equals() and hashCode() have been defined correctly, as has already been stated.
[ December 23, 2008: Message edited by: Peter Lawrey ]

[Songezo Nkukwana]

I re-wrote the initial list "member" to avoid being initialized with duplicated data... So i do not need this function anymore!

Thanks for your inputs

[ramya narayanan]

I'm sure that the original poster hasnot overridden the equals() & hashcode() methods in her collection class.

So when you didn't override the equals() method, it takes the Object's equals() method which compares instances using memory references

so now consider a collection

Set s=new LinkedHashSet();
set.add(car1);//Memory reference 00001234
set.add(car2);//Memory reference 00001235

Eventhough both the references a.k.a car & car1 has a common state like same carno like TN012345 but
since equals() hasnot been overrided in Car class, the equals() from Object class which compares object reference is called.

00001234 != 00001235
car1 != car2
Eventhough both the objects have the same state


So both these objects will be added to set , which by default does not permit duplicates.

Correct me if I'm wrong.
Regards.