Wednesday, April 2, 2008

java Difference between LinkedList ArrayList TreeList:

There are some differences between an java.util.ArrayList and java.util.LinkedList. These differences are not so important for any small applications that processes small amount of data. Then they can very important for any small application that processes huge amount of data. These do affect the performance of a application if not analyzed properly.

An empty LinkedList takes 2-3 times less memory than an empty ArrayList. This becomes important if you plan to keep an adjacency list of nodes for each node in the graph (graph can be like a TreeSet. Tree is a type of graph) and you know beforehand that the nodes will have at most one or two adjacent nodes and the total number of nodes in the graph can be quite large.

Although both ArrayList and LinkedList allow insertion/deletion in the middle through similar operations (ie; by invoking add(int index) or remove(index)), these operations do not offer you the advantage of O(1) insertion/deletion for a LinkedList. For that, you must work through a ListIterator.

The important thing to remember when comparing LinkedList and ArrayList is that linked lists are more faster when inserting and removing at random locations in the list multiple times. If you want to add to the end of the list then an ArrayList would be the best choose.

I will try to supply an answer for the LinkedList vs ArrayList issue. The first thing to do is to look at what interfaces these two implement. This might hint us at their purpose, as Sun usually bound purposes into interfaces.

// lang java
public class ArrayList
extends AbstractList
implements List, RandomAccess, Cloneable, Serializable

public class LinkedList
extends AbstractSequentialList
implements List, Queue, Cloneable, Serializable
We can ignore Cloneable and Serializable as all of the JCF implement those. It’s also quite obvious that both of them implement the List interface, as they need to provide list functionability (backwards iteration, sub listing, etc).
Now for the differences: ArrayList is implementing RandomAccess while LinkedList implementing Queue. You could already tell something about the usage of the two classes.
Now for some implementation notes. The ArrayList is actually encapsulating an actualy Array, an Object[]. When you instanciate ArrayList, an array is created, and when you add values into it, the array changes its size accordingly. This gives you strengths and weaknesses:
• Fast Random Access
You can perform random access without fearing for performence. Calling get(int) will just access the underlying array.
• Adding values might be slow When you don’t know the amount of values the array will contain when you create it, a lot of shifting is going to be done in the memory space when the ArrayList manipulates its internal array.
• Slow manipulation When you’ll want to add a value randomly inside the array, between two already existing values, the array will have to start moving all the values one spot to the right in order to let that happen.
The LinkedList is implemented using nodes linked to each other. Each node contains a previous node link, next node link, and value, which contains the actual data. When new data is inserted, a node is inserted and the links of the surrounding nodes are updated accordingly. When one is removed, the same happens - The surrounding nodes are changing their links and the deleted node is garbage collected. This, as well, gives strengths and weaknesses:
• Fast manipulation As you’d expect, adding and removing new data anywhere in the list is instantanious. Change two links, and you have a new value anywhere you want it.
• No random access Even though the get(int) is still there, it now just iterates the list until it reaches the index you specified. It has some optimizations in order to do that, but that’s basically it.
ArrayList is very useful when a well defined set of data is needed in a List interface as opposed to an array. It can be dynamically changed, but try not to do so frequently throughout the life of the application. LinkedList is there for you to do just that: Manipulating it is very easy, and as long as its used for iteration purposes only and not for random accessing, it’s the best solution. Further, if you need random accessing from time to time, I suggest toArray for that specific moment.
Another point I didn’t raise here is the Queue issue. LinkedList implements extended abilities to the normal List interface which allows it to add and remove elements from its beginning and end. This makes the LinkedList perfect for Queue and Stack purposes - Although in Java 5 they already added a Stack class.