Page 2: Fixed-Length Records

Understanding Fixed-Length Records

When all records in a file have the same size, they're called fixed-length records.

Example: Instructor Table

Let's say we have an instructor table with these fields:

CopyID: 5 characters (5 bytes)
Name: 20 characters (20 bytes) 
Department: 20 characters (20 bytes)
Salary: 8 bytes for numbers
Total: 53 bytes per record

Sample Data:

CopyRecord 0: 10101 | Srinivasan | Comp. Sci. | 65000
Record 1: 12121 | Wu         | Finance    | 90000
Record 2: 15151 | Mozart     | Music      | 40000

Problems with Simple Storage

Problem 1: Block Boundary Crossing

• If block size is 100 bytes and record size is 53 bytes

• Block 1 can hold: Record 0 (bytes 0-52) + part of Record 1 (bytes 53-99)

• Record 1 spans two blocks - need 2 disk reads for 1 record!

Problem 2: Deletion Issues

• When you delete Record 1, what happens to the empty space?

• If we move all records forward, it requires many disk operations

Solutions for Fixed-Length Records

Solution 1: Block-Aligned Allocation

• Only store complete records in each block

• Example: 100-byte block with 53-byte records = 1 record per block (47 bytes unused)

• Wastes space but ensures no record spans blocks

Solution 2: Free List for Deletions Instead of moving records, maintain a list of deleted record positions:

CopyHeader → Points to first deleted record location
Deleted Record 1 → Points to next deleted record
Deleted Record 4 → Points to next deleted record
And so on...

Example of Free List:

CopyFile Header points to position 1
Position 1 (deleted) points to position 4
Position 4 (deleted) points to position 6
Position 6 (deleted) points to NULL

When inserting a new record:

1. Use the position pointed to by header (position 1)

2. Update header to point to next available position (position 4)