1. Data Management and File Organization
Hashing

2. Hashing
Motivation: The number of file access in an indexed file is as many as the tree height (3 or 4 for example)
Hashing method provides a quick access to the records (1 or 2 file access)

3. Definitions
Hash function: A function that returns the location of a record given its key value.
Example: f(25)=1, f(1)=3

4. Definition
Hash table: The data file having the records is called the hash table. Hash table is created using the order returned from the hash function.

5. Example Hash Table Use Key Mod 10 to create the hash table.
Data File Hash Table

6. Collision Problem
The hash function may generate same values for different keys.
Example: Keys 12 and 32 generate same results with hash function :: key mod 10
This is called collision problem

7. Solutions for collision problem
Bucketing: Use buckets as large as n records at each hash table entry
Chaining: Records with the same hash values are chained in a linked list using an overflow area

8. Bucketing

9. Chaining

10. Bucketing Bucket size (n) affects the amount of file I/O
Large bucket size means more I/O to find a record.
Bucketing should be used with chaining for better performance

11. Chaining The chain length is important in I/O speed.
As far as possible, we should keep the chain lengths short.
The performance of the Hashing method depends on choosing a good hash function

12. Questions?

13. Term Project 1 : B-Trees
Assume a file with the following record structure struct Record { int status; int EmployeeID; char Name[30]; char position[10]; float salary; } struct InternalNode int EmpId[4]; InternalNode* links[5]; // Use when pointing to another internal node LeafNode* dlinks[5]; // Use if points to leaf nodes }; Struct LeafNode int Location[4];

14. Term Project 1 : B-Tree
Write a program to create a B-tree index by reading the data file and inserting the key values and the record locations into the index.
Your program should also have a menu to choose the operations by the user.
The possible operations are:
Find a record given Employee ID
Insert a new record (Get the data of the new record from the user)
Delete a record given the Employee ID

15. Term Project 2 : Hashing
Assume a file with the following record structure struct Record { int status; int EmployeeID; char Name[30]; char position[10]; float salary; }

16. Term Project 2 : Hashing
Write a program to read the file and create a hash table.
Use buckets as large as two record size in your hash table.
Assume the number of records is less than 100
Two rightmost digits of each ID can be used as the hash value.
To solve the collision problem use an overflow area at the end of the hash table.

17. Term Project 2 : Hashing
Your program should have a menu with the following options:
Find a record given Employee ID
Delete a record given Employee ID
Insert a new record by reading the record data

18. Assignment
Write a report about choosing hash function for different data types .
Consider integer, float numbers, and character strings as possible data types.