COSC This week. Cloning. Cloning: Be Cautious. Cloning: the wrong type. Cloning: use super.clone()

Transcription

1 This week COSC 3.3 Week 7, February 3, 4. Read chapter 3 Prepare for midterm on Thursday, Feb. 6 9:-: No aids (no calculators, aid sheet, etc) S ome multiple choice, some short answer, and a large programming question Deep/shallow cloning (all of Chapter will be on the midterm) Arrays, array lists, big O notation Cloning Cloning: Be Cautious Pair Pair Triple clone Triple x y x y z Pair // is cloned Triple or Pair? // clone somehow defined in class Pair; do we get the right clone? class Pair //not the standard Pair int x; int y; //override the clone method public Object clone() Pair cloned = new Pair(); cloned.x = this.x; cloned.y = this.y; return cloned; class Triple extends Pair int z; Pair ; public class PairTest public static void main(string[] args) S ystem.out.println( cloned instanceof Triple); //? S ystem.out.println( cloned instanceof Pair); //? //output is false true Cloning: the wrong type Cloning: use super.clone() Pair Triple x x y y z Pair cloned How can this be fixed? x y Pair Triple clone class Pair implements Cloneable int x; int y; //override the clone method public Object clone() throws CloneNotS upportedexception Object cloned = super.clone(); //automatically copies fields return cloned; class Triple extends Pair int z; Pair ; public class PairTest public static void main(string[] args) throws CloneNotSupportedException t.x = ; t.y = ; t.z = 3; S ystem.out.println( cloned instanceof Triple); //? S ystem.out.println( cloned instanceof Pair); //? Triple t = (Triple) cloned; S ystem.out.println( t == t); S ystem.out.println(t.x+","+t.y+","+t.z); //output: true true false,, 3

2 Cloning: so far so good! super.clone() t cloned Pair 3 3 super.clone Triple null null We get the right object. Are all of our problems solved? protected Object clone() throws CloneNotSupportedException Creates and returns a copy of this object. The precise meaning of "copy" may depend on the class of the object. For any object x, the expression: x.clone()!= x will be true, and the expression: x.clone().getclass() == x.getclass() will also be true super.clone() It is typically the case that: x.clone().equals(x) will be true By convention, the returned object should be obtained by calling super.clone. If a class and all of its superclasses (except Object) obey this convention, it will be the case that x.clone().getclass() == x.getclass(). Cloneable interface When a class implements the Cloneable, it tells Object.clone() method that it is legal for that method to make a field-for-field copy of instances of that class. Invoking Object's clone method on an instance that does not implement the Cloneable interface results in the exception CloneNotSupportedException being thrown. Cloneable interface By convention, classes that implement Cloneable interface should override Object.clone This interface does not contain the clone method. Therefore, it is not possible to clone an object merely by virtue of the fact that it implements this interface. Cloneable interface The class Object does not itself implement the interface Cloneable, Hence calling the clone method on an object whose class is Object will result in throwing an exception at run time.

3 But super.clone is Shallow Shallow Cloning class Pair implements Cloneable int x; int y; //override the clone method public Object clone() throws CloneNotS upportedexception Object cloned = super.clone(); return cloned; class Triple extends Pair int z; Pair ; public Triple() = new Pair();.x = ;.y = ; public class PairTest public static void main(string[] args) throws CloneNotSupportedException t.x = ; t.y = ; t.z = 3; S ystem.out.println(((triple) cloned). == ((Triple) t).); Triple t = (Triple) cloned; t..x = 5; t..y = 6; t.z = ; S ystem.out.println(t..x+","+t..y+","+t.z); //output is true 5, 6, 3 t t.x = ; t.y = ; t.z = 3; t 3 cloned 3 t Shallow Cloning t.x = ; t.y = ; t.z = 3; t..x = 5; t..y = 6; t.z = ; t cloned 3 Deep Cloning: Override c lone in Triple class Pair implements Cloneable int x; int y; public Object clone() throws CloneNotSupportedException Object cloned = super.clone(); return cloned; class Triple extends Pair implements Cloneable int z; Pair ; public Triple() = new Pair();.x = ;.y = ; public Object clone() throws CloneNotS upportedexception Object cloned = super.clone(); ((Triple) cloned). = (Pair).clone(); return cloned; public class PairTest public static void main(string[] args) throws CloneNotSupportedException t.x = ; t.y = ; t.z = 3; S ystem.out.println(((triple) cloned). == ((Triple) t).); Triple t = (Triple) cloned; t..x = 5; t..y = 6; t.z = ; S ystem.out.println(t..x+","+t..y+","+t.z); //output is false,, 3 Deep Cloning Collec tion t t.x = ; t.y = ; t.z = 3; t..x = 5; t..y = 6; t.z = ; t cloned 3 When writing programs we often need to work with a group of objects. Examples Concordance: collection of words with their immediate context Phonebook: collection of names with phone numbers Inventory: collection of merchandise and many more.. 3

4 Collec tion S ometimes different types of objects Collec tion S ometimes similar types of objects Operations on Collec tion add, remove, size, get, traverse ArrayLis t Recall material! "" # $"% & //traverse for (int i = ; i < coins.size(); i++) Coin c = (Coin)coins.get(i); do some thing with c Adding/Removing Elements overwrite an existing value coins.set(4, anickel); add a new value before the index add(i, c) remove removes an element at an index 4

5 Storing Numbers in an Array List Array lists store objects Use wrapper classes to store numbers '%(" )) '%("*++, %(" %)) ))%("-"% )) %)) '%("%("-"% Also have. and /" wrappers Construct array: %(" Arrays Store in variable of type double[] %(" %(" Arrays Arrays have fixe d le ngth Arrays have element of specific type, not!( Use [] to access element: 3 *++, Get array length as "4. (Not a method!) Copying Arrays Copying an array reference yields a second reference to the same array %(" %(" "" %(" ) Use " to make true copy %(" ) %("" Copying Array Elements S ystem.arraycopy(from, froms tart, to, tos tart, count); 5

6 Adding/Removing Array Elements Add an element: 5) ) 6 "4 & & # Remove an element: 5) ) 6 "4 & & Partially Filled Arrays Partially Filled Arrays Array length = maximum number of elements in array Usually, array is partially filled Need companion variable to keep track of current size Uniform naming convention: final int DATA_LENGTH = ; double[] data = new double[data_length]; int datasize = ; Update datasize as array is filled: data[datasize] = x; datasize++; Partially Filled Arrays Remember to stop at 5 when looking at array elements: % % 6 Be careful not to overfill the array 5 9 "4 8 85%)"5&& %"" Be careful to grow the array when needed: %(" ' %("* : "4 5) ' "4 ' Growing an Array 6

7 Cos t of Operations Computer programs use computer resources in order to carry out their tasks Two important resources: CPU Memory (primary storage) Other resources include secondary storage, printers, bus bandwidth, etc. Cos t of Operations We are mainly interested in CPU usage (running time) and memory usage of a program. Why should we be concerned with these? If your program is too slow or uses too much memory, then nobody will use it. Cos t of Operations Two methods for searching for a phone number in a telephone book: Linear search: start from the first page; flip through every page and compare every name with the key until found Binary search: open the book in the middle; compare key with the name in the middle; if key is smaller then do a binary search on first half; otherwise do a binary search on second half Cos t of Operations Assume the telephone book contains,, entries. How many lookups does each take: Linear search:,, lookups Binary search: each iteration cuts the number of names by a factor of until one is left. Hence, if we do this x times, then we seek x so that,,/ x is roughly. This makes x to be about = log (,,) Time Complexity Even though both algorithms achieve the same results; one uses much more resources than the other one. In general we express the running time of a program as a function of its input size: n Time complexity of sorting n integers Time complexity of searching for a word in a dictionary As n grows so does the running time Cos t of Operations Note that you cannot do a binary search on a collection whose entries are not sorted Also regardless of whether entries are sorted or not, you cannot do binary search on a collection of items which only allows sequential access (e.g. linked list) 7

8 Random Access Cost of access is independent of location; Get content of cell 6 Sequential Access Cost depends on location. Get the object after/before the lion or a given position Time Complexity We say a program takes f(n) operations (basic operations like add, subtract multiply, assign) to perform its task S ome examples f(n) =,, n f(n) = n + n + f(n) = n log(n) Time and Spac e Complexity Big O notation. A function f(n) is O(g(n)) (f is of order g) iff there exists two positive constants c and n such that f(n) < c g(n) For all n >= n Big-O Notation A function f(n) is said to be Big-O of g(n), written f(n) O(g(n)) c g(n) f(n) Implementing Collec tions S upported operations: Creation add(object) - add at the end of the list remove(index) - remove object at index size() - get the size of the list get(index) - get the object at index n n 8

9 Implementing Collec tions We can realize collections using different underlying structures: arrays linked lists trees hash tables We will investigate arrays and linked for now MyArrayList: A wrapper for an array public class MyArrayList private Object[] objs; private int last; public MyArrayList(int max) objs = new Object[max]; last = ; public void add(object o) objs[last++] = o; public void remove(int index) for(int i=index;i<last-;i++) objs[i] = objs[i+]; last--; public Object get(int index) return objs[index]; public int size() return last; Best, worst, average remove element given index i Best case: i == last Worst case: i == Average case: size (cost of removing i)/size i = the amount of work is proportional to (n + (n -) )/ n = n(n + ) / (n) = (n+)/ Best, worst, average Best case performance Often easy to compute (quickest way through the code). Worst case performance Again usually easy to compute (slowest path through the code). Average case performance Can be difficult to define average input. Back to MyArrayList Analysis often focuses on the worst case Does this make sense? Performance requirements often ask for running time better than x Analysis of MyArrayList Add is ; Very cheap Remove is : worst case Cos t of operations Add/get/size - seem very efficient Independent of the size of collection Remove - possibly move up to n elements to remove the item at element index How can we categorize the cost of different operations? 9

10 Sample s - not proportional to n (constant time) Accessing an element of an array - proportional to the size of the problem S earching through a list (remove operation) O(n ) - quadratic in n S etting elements in a matrix O(log n) - order log n S earching an ordered list via binary search O(n log n) - order n log n S orting a list (efficiently) O( n ) - exponential Sample s Problem is intractable for reasonably large problems. Many interesting problems are O( n ) (or worse) 5 9 Traveling S alesman problem Simple c omputations n 3 + n is O(n 3 ) 5n 3 + n is O( n ) n logn + n is O(n ) Comparis on of func tions Let n = 6 and assume every operation takes -8 seconds f(n) = n time = 6* -8 seconds f(n) = n time = 36* -6 seconds f(n) = n 3 time = 6* -5 seconds f(n) = n 4 time.3 seconds f(n) = n 5 time.3 minutes f(n) = n time 37 years As ymptotic Analys is Given a problem, how do you compute? Decide what the unit cost is Array access, comparison, numerical operation, Only concerned with performance for large n (n >= n ) Find the dominant term log n, n, nlog n, n, n 3,.. n How do we make ArrayList grow? public MyArrayList(int max) objs = new Object[max]; last = ; this.max = max; public void add(object o) objs[last++] = o; if(last == objs.length) Object[] temp = new Object[objs.length + max]; for(int i=;i<objs.length;i++) temp[i] = objs[i]; objs = temp;

11 Performanc e of MyArrayLis t Add is now (potentially) quite expensive Add is Remove is Get is S et(index, o) is MyArrayList vs. ArrayList Now quite similar Many more methods in ArrayList, but their implementations are quite straightforward Will write one more here indexof(object o) public int indexof(object o) for(int i=;i<last;i++) if(objs[i].equals(o)) return i; return -; boolean equals(objec t o) if(o == null) return false; return o == this Define it more intelligently for your classes. Can we improve on MyArrayLis t? S uppose instead of using an array to hold the collection, we use something else (like a linked list) Implement add, remove, get, set, indexof Obvious linked lis t implementation S ingly linked list (much like the example we did before). Two classes, a node class (MyNode) A linkedlist class (MyLinkedList) public class MyNode private Object o; private MyNode next; public MyNode(Object o) this.o = o; this.next = null; public Object geto() return o; public void seto(object o) this.o = o; public void setnext(mynode next) this.next = next; public MyNode getnext() return next;

12 public class MyLinkedList private MyNode head; public MyLinkedList() head = null; public void add(object o) MyNode tmp = new MyNode(o); MyNode pre = null; for(mynode cur=head; cur!= null; pre=cur, cur=cur.getnext()) ; if(pre == null) head = tmp; else pre.setnext(tmp); public int indexof(object o) int i = ; for(mynode cur=head; cur!= null; cur=cur.getnext(), i++) if(cur.geto().equals(o)) return i; return -; public void remove(int index) MyNode pre = null; MyNode cur = head; for(int i=;i<index;i++) pre = cur; cur = cur.getnext(); if(pre == null) head = head.getnext(); else pre.setnext(cur.getnext()); public Object get(int index) MyNode cur=head; for(int i=;i<index;i++) cur = cur.getnext(); return cur.geto(); public int size() int last = ; for(mynode cur=head; cur!= null; cur=cur.getnext(), last++); ; return last; public void set(int index, Object o) MyNode cur=head; for(int i=;i<index;i++) cur = cur.getnext(); cur.seto(o); So how did we do? Can we improve things? Create add remove get MyArrayList MyLinkedList One quick improvement Maintain the length of the list in the class S et to initially Increment by in add Decrement by in delete Return its value in size set indexof size

13 Revised version But can we do better? Create add remove MyArrayList MyLinkedList remove, get set and indexof require search Unlikely that we can improve on these with the current data structure Can we improve on add? get set indexof size Add spends all of its effort finding the end of the list. Can we maintain a last node? Add, remove will need some work public class MyLinkedList private MyNode head, tail; public MyLinkedList() head = null; tail = null; public void add(object o) MyNode tmp = new MyNode(o); if(head == null) head = tmp; tail = tmp; else tail.setnext(tmp); tail = tmp; public void remove(int index) MyNode pre = null; MyNode cur = head; for(int i=;i<index;i++) pre = cur; cur = cur.getnext(); if(pre == null) head = head.getnext(); if(head == null) tail = null; else pre.setnext(cur.getnext()); if(cur.getnext() == null) tail = pre; Revised version So how bad is this Create add remove get set indexof MyArrayList MyLinkedList S uppose all VIS A card numbers (5 digits) are in a list ( 5 numbers) S uppose each unit operation can be done in -7 seconds. Then it will only take on the order of 8 seconds to search list (3 years). For big problems, is not good enough size 3

14 But for small problems ArrayList Vector (use ArrayList instead) LinkedList Great for S tack and Queue implementations (more on these later) What makes lists? Being able to access arbitrary elements. If we limit access to the list, can we enhance performance? Two classic alternatives Limit add to one end of the list Limit remove to one end of the list LIFO (stack) FIFO (queue) Create Stack Push - add an element to the top of the stack Pop - remove the top element of the stack isempty - check if the stack is empty Create Queue Enqueue - add to the end of the queue Dequeue - remote top element from the queue isempty - check to see if the queue is empty Two-Dimens ional Arrays Matrix with rows and columns Example: Tic Tac Toe board char[][] board = new char[3][3]; board[i][j] = 'x'; File Tic Tac Toe.java /** A 3 x 3 Tic-Tac-Toe board. 3 */ 4 public class TicTacToe 5 6 /** 7 Constructs an empty board. 8 */ 9 public TicTacToe() board = new char[rows ][COLUMNS ]; 3 //fill with spaces 4 for (int i = ; i < ROWS ; i++) 5 for (int j = ; j < COLUMNS ; j++) 6 board[i][j] = ''; 7 4

15 8 9 /** Sets a field in the board. The field must be param i the row param j the column index param player the player ('x'or 'o') 4 */ 5 public void set(int i, int j, char player) 6 7 if (board[i][j]!= '') 8 throw new IllegalArgumentException("Position occupied"); 9 board[i][j] = player; /** 33 Creates a string representation of the board such as 34 x o 35 x 36 o return the string representation 38 public S tring tos tring() 39 4 S tring r = ""; 4 for (int i = ; i < ROWS ; i++) 4 43 r = r + " "; 44 for (int j = ; j < COLUMNS ; j++) 45 r = r + board[i][j]; 46 r = r + " \n"; return r; private char[][] board; 5 private static final int ROWS = 3; 53 private static final int COLUMNS = 3; 54 File Tic Tac ToeTes t.java import javax.swing.j OptionPane; 3 /** 4 This program tests the TicTacToe class by prompting the 5 user to set positions on the board and printing out the 6 result. 7 */ 8 public class TicTacToeTest 9 public static void main(string[] args) char player = 'x'; 3 TicTacToe game = new TicTacToe(); 4 while (true) 5 6 System.out.println(game); //calls game.tostring() 7 String input = JOptionPane.showInputDialog( 8 "Row for " + player + " (Cancel to exit)"); 9 if (input == null) System.exit(); int row = Integer.parseInt(input); input = J OptionPane.showInputDialog( "Column for " + player); 3 int column = Integer.parseInt(input); 4 game.set(row, column, player); 5 if (player == 'x') player = 'o'; else player = 'x'; Read chapter 3 This week Prepare for midterm this week 7:-9: No aids (no calculators, cheat sheets, etc) S ome multiple choice questions, some short answer questions, and a larger question Arrays, array lists, big O notation 5