Queues Data Structures and Design with Java and

Queues Data Structures and Design with Java and JUnit © Rick Mercer 18 -1

A Queue ADT First in first out access w A Queue is another name for waiting line w Example: Submit jobs to a network printer ¾ What gets printed first? · · Print the least recently submitted no priorities given Add new print jobs at the end of the queue w Queues provide First In First Out (FIFO) access to elements could also say Last In Last Out (LILO) 18 -2

The Java Queue class? w Some languages have a Queue class or queue is part of a library that works with the language ¾ Java 1. 4 used class Linked. List to offer FIFO functionality by adding methods add. Last(Object) and Object(get. First) ¾ Java 1. 5 added a Queue interface and several collection classes: Array. Blocking. Queue<E> and Linked. Blocking. Queue<E> w Outline of what we'll do ¾ ¾ Specify a Queue ADT as a Java interface Show a difficult to implement array based implementation 18 -3

Designing a Queue Interface w Queues typically provide these operations ¾ ¾ adds an element at the end of the queue peek returns a reference to the element at the front of the queue removes the element from the front of the queue and returns a reference to the front element is. Empty returns false if there is at least one element in the queue 18 -4

Specify an interface w We will use an interface to describe a queue ADT ¾ The interface specifies method names, return types, the type of elements to add, and hopefully comments w interface Our. Queue declares we must be able to add and remove any type of element ¾ Collection class must have <E> to make it a generic type 18 -5

Interface to specify a FIFO Queue import java. util. No. Such. Element. Exception; public interface Our. Queue<E> { // Return true if this queue has 0 elements public boolean is. Empty(); // Store a reference to any object at the end public void add(E new. El); // Return a reference to the object at the // front of this queue public E peek() throws No. Such. Element. Exception; // Remove the reference to the element at the front public E remove() throws No. Such. Element. Exception; } 18 -6

Let Slow. Queue implement the Queue interface w We need to store an Object[] ¾ an array of Object objects avoids having queue of int and people and cars, and. . . public class Slow. Queue<E> implements Our. Queue<E> { private int back; private Object[] data; //. . . w Now implement all methods of the Our. Queue interface as they are written ¾ plus a constructor with the proper name 18 -7

Bad array type queue w Queue as an array could have ¾ the front of the queue is always in data [0] public Slow. Queue(int max) { data = new Object[max]; back = -1; } data[0] null back == -1 data[1] data[2] null data[3] null So far so good. An empty queue 18 -8

First version of add public void add(E element) { // This method will be changed later back++; data[back] = element; } w Send an add message a. Queue. add("a"); data[0] "a" back == 0 data[1] null data[2] data[3] null So far so good. A queue of size 1 18 -9

add another public void add(E element) { // This method will be changed later back++; data[back] = element; } w Send two more add messages a. Queue. add("b"); a. Queue. add("c"); data[0] "a" back == 2 data[1] "b" data[2] data[3] "c" null So far so good. A Queue of size 3 18 -10

Array Implementation of a Queue w During remove, slide all elements left if size were 999, then 998 assignments would be necessary Before remove "a" "b" "c" null back A poor remove algorithm After remove "b" "c" back null 18 -11

Effect of queue operation using an array with a "floating" front "a" add("a") add("b") add("c") front remove() "a" "b" "b" "c" "a" "b" ? back "c" "d" back front remove() ? back front add("d") "c" "d" front back 18 -12

What happens next when back equals array. length? add("e") "a" "b" "c" "d" front back indexes the last array index w However, this queue is not full w Where do you place "e"? w back "e" data[0] is available ¾ What code would increment back correctly even when back is referencing the last available array index ¾ ? ________________ ? 18 -13

The Circular Queue w A "circular queue" implementation uses wraparound The queue has "c" "d" "e" ¾ either increase back by 1 or set back to 0 "d" "e" "a" front "c" "b" back data[0] add("e") now works in this "circular" queue. It reuses previously used array indexes 18 -14

Implementing a Circular Queue w Still have to work with arrays, not circles ¾ In order for the first and last indices to work in a circular manner: · · ¾ increase by one element at a time after largest index in the array, go to zero. back = 0 1 2 3 0. . . could contain code you just wrote on previous slide w But what is an empty queue? ¾ What values should be given to front and back when the queue is constructed? 18 -15

Problem: A full queue can not be distinguished from an empty queue One option is to have the constructor place back one index before front then increment back during add back front a empty a 1 in q front back 3 in q c a d full q b c b back What does back == front imply? An empty or full queue? 18 -16

Corrected Circular Queue w Use this trick to distinguish between full and empty queues ¾ The element referenced by front never indexes the front element— the “real” front is located at next. Index(front) private int next. Index(int index) { // Return an int to indicate next position return (index + 1) % data. length; } ¾ For example, use this during peek() return data[next. Index(front)]; 18 -17

Correct Circular Queue Implementation Illustrated back front back "a" empty 1 in q front "a" 2 in q "b" back full q "c" "b" back The front index is always 1 behind the actual front This wastes one array element but it's no big deal 18 -18

Correct Circular remove Implementation Illustrated front "a" full q "c" back 2 in q "b" "c" back "b" 1 in q empty q "c" back front remove three times to make this queue empty front 18 -19