1. Assignment 2: A Simple Address Book Andy Wang Data Structures, Algorithms, and Generic Programming

2. Learning Objectives  Gain further experience in C++  Manage project develop using make  Apply the concept of hash functions

3. Mission  Write a simple address book  Use hash functions to access a table Insertion Lookup Deletion

4. Deliverables (Due 9/26/2003)  makefile  hashtable.h  hashtable.cpp  main.cpp  Development logs

5. More on Development Log  Due in each class  Cumulative 9/17, turn in a log that covers 9/15 – 9/17 9/22, turn in a log that covers 9/15 – 9/22 And so on…

6. Requirements  Create a proj2 directory makefile hashtable.h hashtable.cpp main.cpp

7. hashtable.h  Interface of the class HashTable  Required public interface: bool Insert(const string &name, const string &addr); bool Lookup(const string &name, string *addr) const; string Remove(const string &name);

8. Constructor  Hashtable(unsigned int size);  Default size: 5

9. Insert() bool Insert(const string &name, const string &addr); Return value true on success false on failure Duplicate name Table is full

10. Insert() Name: Michael Jackson Address: 0 1 2 3 Hash Table

11. Insert() Name: Michael Jackson Address: Never-Never Land 0 1 2 3 Hash(name) Hash Table

12. Insert() Name: Michael Jackson Address: Never-Never Land 0 1Michael JacksonNever-Never Land 2 3 Hash(name) Hash Table

13. Insert() Name: Mickey Mouse Address: Disneyland 0 1Michael JacksonNever-Never Land 2 3 Hash Table

14. Insert() Name: Mickey Mouse Address: Disneyland 0 1Michael JacksonNever-Never Land 2 3 Hash(name) Hash Table Oh, Boy!

15. One problem with hashing—collision  Collision handling techniques: Increase the hash table size Chaining Linear probing Quadratic probing Double hashing

16. Increase the Hash Table Size  If collision, double the size, rehash all entries + Conceptually simple - Unbounded growth of table size Birthday Paradox In a group of 60 people, you are very likely to find two people with the same birthday. The probability of collisions is higher than you think!

17. Chaining  Requires a data structure from a later lecture  Basic idea: each table entry is associated with multiple (key, value) pairs  Sequentially look through those pairs + incremental growth of the table - poor performance when an entry is associated with too many (key, value) pairs

18. Linear Probing  Sequentially find the next empty table entry (Hash(key) + 1) % size (Hash(key) + 2) % size … + Simple + Use all table entries before size increase - Clustering (nonuniform distribution of table entries) - Can degenerate into sequential searches

19. Quadratic Probing  Try to avoid clustering (Hash(key) + 1 2 ) % size (Hash(key) + 2 2 ) % size… + Simple - Secondary clustering (not as severe)

20. Double Hashing  Use another hash function to determine the skipping distance If Hash 1 (key) points to an occupied entry Use (Hash 1 (key) + Hash 2 (Hash 1 (key))) % size (Hash 1 (key) + 2*Hash 2 (Hash 1 (key)) % size… + Avoids secondary clustering + Widely used in compilers

21. Lookup() bool Lookup(const string &name, string *address) const; Return value true if found address contains the address false if not found

22. Lookup() Name: Michael Jackson 0 1Michael JacksonNever-Never Land 2 3 Hash Table

23. Lookup() 0 1Michael JacksonNever-Never Land 2 3 Hash(name) Hash Table Name: Michael Jackson

24. Remove() string Remove(const string &name) const; Return value The address associated with the name

25. Remove() Name: Michael Jackson 0 1Michael JacksonNever-Never Land 2 3 Hash Table

26. Remove() 0 1Michael JacksonNever-Never Land 2 3 Hash(name) Hash Table Name: Michael Jackson

27. Remove() 0 1 2 3 Hash(name) Hash Table Name: Michael Jackson

28. Your Hash Function  Hash(name) // may contain spaces  Develop/choose your own hash function(s) Use all characters in the name Use the ordering information of each character Short names should not cluster in the table

29. Milestones 1. Create all files Add #include into.cpp files Add #ifndef into.h files Add files names into the makefile make clean make

30. Milestones 2. Add the class definitions to the header files Try to compile 3. Write the skeleton code in your.cpp files Nothing but empty functions Try to get the interface to compile 4. Write pseudo code (comments) in each skeleton function

31. Milestones 5. Implement the constructor/destructor routines Add private functions as needed Compile and test those routines 6. Implement a dump routine to visualize the hash table

32. Milestones 7. Implement the accessor routines that only read your private data members Compile and test those routines 8. Implement the accessor routines that write to your private data members Compile and test those routines

33. Milestones 9. Implement the Insert routine Implement the associated hash function(s) Test the hash function in isolation Check for clustering Check for duplicate names 10. Implement collision handling

34. Milestones 11. Implement the Lookup routine Compile and test the routine 12. Implement the Remove routine. Compile and test the routine

35. Pace Yourself  Try to accomplish 1 to 2 milestones per day  Start your project early!