A good Sorted List for Java

Depending on how you're using the list, it may be worth it to use a TreeSet and then use the toArray() method at the end. I had a case where I needed a sorted list, and I found that the TreeSet + toArray() was much faster than adding to an array and merge sorting at the end.

Phuong:

Sorting 40,000 random numbers:

0.022 seconds

import java.util.ArrayList;
import java.util.Collections;
import java.util.List;
import java.util.Random;


public class test
{
    public static void main(String[] args)
    {
        List<Integer> nums = new ArrayList<Integer>();
        Random rand = new Random();
        for( int i = 0; i < 40000; i++ )
        {
            nums.add( rand.nextInt(Integer.MAX_VALUE) );
        }

        long start = System.nanoTime();
        Collections.sort(nums);
        long end = System.nanoTime();

        System.out.println((end-start)/1e9);
    }
}   

Since you rarely need sorting, as per your problem statement, this is probably more efficient than it needs to be.

This is the SortedList implementation I am using. Maybe this helps with your problem:

import java.util.Collection;
import java.util.Collections;
import java.util.Comparator;
import java.util.LinkedList;
/**
 * This class is a List implementation which sorts the elements using the
 * comparator specified when constructing a new instance.
 * 
 * @param <T>
 */
public class SortedList<T> extends ArrayList<T> {
    /**
     * Needed for serialization.
     */
    private static final long serialVersionUID = 1L;
    /**
     * Comparator used to sort the list.
     */
    private Comparator<? super T> comparator = null;
    /**
     * Construct a new instance with the list elements sorted in their
     * {@link java.lang.Comparable} natural ordering.
     */
    public SortedList() {
    }
    /**
     * Construct a new instance using the given comparator.
     * 
     * @param comparator
     */
    public SortedList(Comparator<? super T> comparator) {
        this.comparator = comparator;
    }
    /**
     * Construct a new instance containing the elements of the specified
     * collection with the list elements sorted in their
     * {@link java.lang.Comparable} natural ordering.
     * 
     * @param collection
     */
    public SortedList(Collection<? extends T> collection) {
        addAll(collection);
    }
    /**
     * Construct a new instance containing the elements of the specified
     * collection with the list elements sorted using the given comparator.
     * 
     * @param collection
     * @param comparator
     */
    public SortedList(Collection<? extends T> collection, Comparator<? super T> comparator) {
        this(comparator);
        addAll(collection);
    }
    /**
     * Add a new entry to the list. The insertion point is calculated using the
     * comparator.
     * 
     * @param paramT
     * @return <code>true</code> if this collection changed as a result of the call.
     */
    @Override
    public boolean add(T paramT) {
        int initialSize = this.size();
        // Retrieves the position of an existing, equal element or the 
        // insertion position for new elements (negative).
        int insertionPoint = Collections.binarySearch(this, paramT, comparator);
        super.add((insertionPoint > -1) ? insertionPoint : (-insertionPoint) - 1, paramT);
        return (this.size() != initialSize);
    }
    /**
     * Adds all elements in the specified collection to the list. Each element
     * will be inserted at the correct position to keep the list sorted.
     * 
     * @param paramCollection
     * @return <code>true</code> if this collection changed as a result of the call.
     */
    @Override
    public boolean addAll(Collection<? extends T> paramCollection) {
        boolean result = false;
        if (paramCollection.size() > 4) {
            result = super.addAll(paramCollection);
            Collections.sort(this, comparator);
        }
        else {
            for (T paramT:paramCollection) {
                result |= add(paramT);
            }
        }
        return result;
    }
    /**
     * Check, if this list contains the given Element. This is faster than the
     * {@link #contains(Object)} method, since it is based on binary search.
     * 
     * @param paramT
     * @return <code>true</code>, if the element is contained in this list;
     * <code>false</code>, otherwise.
     */
    public boolean containsElement(T paramT) {
        return (Collections.binarySearch(this, paramT, comparator) > -1);
    }
    /**
     * @return The comparator used for sorting this list.
     */
    public Comparator<? super T> getComparator() {
        return comparator;
    }
    /**
     * Assign a new comparator and sort the list using this new comparator.
     * 
     * @param comparator
     */
    public void setComparator(Comparator<? super T> comparator) {
        this.comparator = comparator;
        Collections.sort(this, comparator);
    }
}

This solution is very flexible and uses existing Java functions:

• Completely based on generics
• Uses java.util.Collections for finding and inserting list elements
• Option to use a custom Comparator for list sorting

Some notes:

• This sorted list is not synchronized since it inherits from java.util.ArrayList. Use Collections.synchronizedList if you need this (refer to the Java documentation for java.util.ArrayList for details).
• The initial solution was based on java.util.LinkedList. For better performance, specifically for finding the insertion point (Logan's comment) and quicker get operations (https://dzone.com/articles/arraylist-vs-linkedlist-vs), this has been changed to java.util.ArrayList.

It seems that you want a list structure with very fast removal and random access by index (not by key) times. An ArrayList gives you the latter and a HashMap or TreeMap give you the former.

There is one structure in Apache Commons Collections that may be what you are looking for, the TreeList. The JavaDoc specifies that it is optimized for quick insertion and removal at any index in the list. If you also need generics though, this will not help you.