Question

I’ve something like this

Object[] myObjects = ...(initialized in some way)...
int[] elemToRemove = new int[]{3,4,6,8,...}

What’s the most efficient way of removing the elements of index position 3,4,6,8… from myObjects ?

I’d like to implement an efficient Utility method with a signature like

public Object[] removeElements(Object[] object, int[] elementsToRemove) {...}

The Object[] that is returned should be a new Object of size myObjects.length – elemToRemove.length

Answer 1

You will need to create a new array, since arrays do not change in size (you cannot “remove” elements while keeping the same array, it would create holes somewhere). I suggest this:

Object[] nobjs = Arrays.copyOf(myObjects, myObjects.length - elemToRemove.length);
for (int i = 0, j = 0, k = 0; i < myObjects.length; i ++) {
    if (j < elemToRemove.length && i == elemToRemove[j]) {
        j ++;
    } else {
        nobjs[k ++] = myObjects[i];
    }
}

This assumes that elemToRemove is already sorted, with no duplicates, and contains only indices valid within the source array. The Arrays.copyOf() call is used only to make sure that the runtime type of the new array is identical to that of the source array.

If you often have to “remove” data from arrays, thus requiring the creation of new arrays, then Java arrays might not be the most efficient data structure for you. LinkedList or ArrayList may be more appropriate.