Assignment 2

1
Assignment 2

• Objective Design and write a program to simulate
  the operations in a file processing system
• Programming environment
• OS UNIX
• Program language C/C
• Command line format ass2 inputfile
• Sample inputs will be provided for testing your
  program.
• Warning Plagiarism

2
Assignment 2

• Submitted materials
• studentID_ass2.tar sent to the course account
  including
• Makefile
• Source files ass2.c etc.
• Design report (short and well-written)
• Name, student ID and email address.
• Platform / Compiler
• File structure
• Data structure
• Algorithms
• Due Date Feb. 25th, 2001
• Late submission policy 20 reduction per day

3
Assumptions

• There can be multiple files in the system. Each
  file consists of a collection of blocks. The size
  of each block, BlockSize, is equal to 1KBytes.
• A file contains a set of fixed-length records.
  The record length, RecordSize, varies for
  different files though fixed in one file. There
  is no record cross multiple blocks, i.e.
  RecordSize
• Each record is made up of two fields, key and
  data. key is unique for each record in a file.
  The lengths of the two fields are KeySize and
  DataSize respectively. KeySize is assumed to be
  10Bytes. RecordSizeKeysizeDataSize

4
Design Hints

• Please note that the data transfer unit is a
  block. read() or write() one block a time from or
  to the disk.
• All the blocks are of the same size. To locate
  the starting point of a block identified by
  blockID in a file, use lseek() system call to
  move the file pointer appropriately.
• The first block in a file is the header block
  which stores the structural information of the
  file, such as block size, block number, record
  size, record number, sorting tag, the first block
  ID etc. ( the header block structure )
• All the data blocks in a file are organized in a
  doubly linked list. ( the data block structure )

5
Design Hints---Cont.

• An array in the main memory can be used to keep
  the control information of all the opened files.
• Bear the following operations in mind when
  choosing from different strategies.
• Use low level I/O (unbuffered)in C. Such system
  calls in UNIX are
• open(), close()
• read(), write()
• lseek(), rewind()
• remove(), unlink(),
• etc.

6
File Operations

• CREATE filename recsize -s
• create a new file. Report error if the file
  already exists.
• filename the name of the file to be created.
• recsize the record size in the file in term of
  number of bytes. Recall that the KeySize10, so
  the DataSizerecsize-10.
• -s is optional. When specified, it indicates that
  the file is sorted in ascending order of the key.
  Otherwise the file is not sorted.
• The newly created file consists of only the
  header block initially.

7
File Operations - Cont.

• SELECT reckey FROM filename
• retrieve one record from a file.
• filename the name of the file to be selected
  from.
• reckey the key of the record to be retrieved.
• Open the file first if it has not been opened. If
  the file does not exist, report an error.
• Scan all the blocks in the file. If the record
  identified by the given key exists, display the
  data field of the record, otherwise show the
  message of nonexistence.

8
File Operations - Cont.

• INSERT reckey recdata INTO filename
• insert one record into a file
• filename the name of the file to be inserted
  into.
• reckey key of the record to be inserted
• recdata data of the record to be inserted.
• Open the file first if the file has not be
  opened. If the file does not exist, report an
  error.
• Report error if a record with the same key
  already exists in the file since duplicate keys
  are not allowed.
• Insert the record properly (keep the file sorted
  if required) whether to a new block or an empty
  slot on an existing block of the file.

9
File Operations - Cont.

• DELETE reckey FROM filename
• delete one record from a file
• filename name of the file to be deleted from
• reckey key of the record to be deleted
• Open the file first if the file has not be
  opened. If the file does not exist, report an
  error.
• Report an error if the record with the given key
  does not exist

10
File Operations - Cont.

• DROP filename
• remove a file
• filename the name of file to be removed
• The file should be closed first.
• Report an error if the file does not exist
• ANALYZE filename
• display the information on all the blocks (header
  block and data block) of a file
• filename the name of the file
• Open the file first if it has not been opened.
  Read each block of the file and show the content
  of the block in a readable way.