Overview

In this lab you will implement a program that reads entries from a roster file and inserts them into an extendible hash table. Your program will then request last names from the user and print out all roster entries with that last name.

A sample entry from the roster might look as follows:

Riggs       Gerald    6.0  TB Sr Tennessee Chattanooga, TN     

1. last name (12 chars)
2. first name (10 chars)
3. height (5 chars)
4. position (3 chars)
5. rank (3 chars)
6. team (10 chars)
7. home town (20 chars)

You may make the following assumptions:

1. no players have the same first and last name.
2. first and last names are single words.
3. the code provided to you will read the roster file and store it into a Roster_Record class object. Look at the class declaration for Roster_Record in hsearch.h. You will notice that the class declaration gives you methods for creating a Roster_Record, printing a Roster_Record, and accessing the fields in a Roster_Record. You will also notice that the hsearch.cpp file contains code for reading the records from a file.

Sample binary code is available as /home/parker/courses/cs302/lab7/hsearch. You can run this program as follows:

 hsearch roster 

Enter last name (Ctrl-D to exit): Wilson 

Ellix Wilson 5.10 LB So Tennessee Memphis, TN
Nick Wilson 6.2 LB Sr Tennessee Knoxville, TN

Enter last name (Ctrl-D to exit): Manning

Peyton Manning 6.5 QB 18 Colts New Orleans, LA
Eli Manning 6.4 QB 10 Giants New Orleans, LA

Enter last name (Ctrl-D to exit): Haralson

Parys Haralson 6.2 DE Sr Tennessee Flora, MS

Enter last name (Ctrl-D to exit): Jones

Couldn't find Jones in the roster

Enter last name (Ctrl-D to exit): ^D

Details

Implement an extendible hashing scheme that:

1. reads in the roster file and uses extendible hashing to store the roster file on the simulated disks provided by the disks package, and

2. allows searches on the hashed file using last names. Your program should print out all roster entries with the given last name.

To assist you in getting started, we have provided you with an include file, /home/parker/courses/cs302/labs/lab7/hsearch.h, and some initial code, /home/parker/courses/cs302/labs/lab7/hsearch.cpp. You can complete this lab by editing hsearch.cpp in the following manner:

1. Write the function split_rec_node. Peruse the hsearch.cpp file to figure out what this function should do.

2. Write the function search. Peruse the hsearch.cpp file to figure out what this function should do.

3. Write the code to insert the records in the roster file into the extendible hash table.

4. Write the code to search for records requested by the user.

Information that May Prove Helpful

1. To simplify this lab, keep the index for the extendible hash table in main memory rather than on disk. The data structure will be an array of pointers to index records. You will start with an index of two entries and double it as necessary.

2. In each record page, records are stored in alphabetical order by last name.

3. You can find disk.h in /home/parker/courses/cs302/include and disk.o in /home/parker/courses/cs302/objs. If you want to try to run on your home computer you can find the code for disk.cpp in /home/parker/courses/cs302/src/disk.cpp.

4. Disks have the following parameters:
   • Each disk contains a maximum of 4 pages.
   • Each record page contains a maximum of 5 records.

Helpful Functions for Your Lab

• hash: hash takes a pointer to a character string (a last name in this lab) and the number of leading bits in the hash table. hash returns a hash key with the desired number of bits.

• assign_disk_space: assign_disk_space takes a reference to a page buffer, finds a free page on a disk, and assigns the page to the page buffer (i.e., it sets the disk and page_num fields of the page buffer to the appropriate values). You should use assign_disk_space whenever you want to get a free page on a disk (assign_disk_space is just like malloc, but it works for external memory).

• EXTENDIBLE_HASH_TABLE constructor: The constructor initializes the hash table to a size of 2. It creates two index entries and two empty pages for records, writes the two pages to disk, and makes the index entries point to the pages. You should peruse the code in this function to get some ideas about how your code will be written. You should pay particular attention to how assign_disk_space is used and to how the disk addresses of the pages are extracted from the page buffer and assigned to the index entries.

• housekeeping: this function defines variables that assign_disk_space needs, including a disk array. You should not have to worry about this function.

What to Submit

1. A makefile
2. roster
3. hsearch.cpp
4. hsearch.h
5. README (optional): If you cannot complete the lab then submit a README file that tells the TAs what you were able to accomplish. If possible include some debugging statements that shows what happens as you insert records into the hash key. Appropriate debugging information would be a statement telling the user which roster entry is being inserted, what the page looks like once the roster entry is inserted, and what the index looks like after each insertion.