Chapter 11: Sequential File Organization

1
Chapter 11Sequential File Organization
2
What is Sequential Organized File?

• In sequential organized file, the records are
  written consecutively when the file is created
  and must be accessed consecutively when the file
  is later used for input.
• In many cases the records of a sequential file
  are ordered by the value of some field in each
  record. Such file is said to be a sorted file
  the field whose value is used to determine the
  ordering is known as the sort key.
• A file may be sorted in ascending or descending
  order by a sort key comprised of one or more
  fields.

3
Structure of a Sequence File
Beginning of file
End of file
4
Ordering Records

• Processing requirement determine the appropriate
  order for sorting the records of a sequential
  file
• For example, the records of a sequential file
  used to generate the white pages of telephone
  directory would be sorted by subscriber last name
  or first name. However, in some cases, it may be
  useful to have a telephone directory with records
  sorted by subscriber address, or by zip code, or
  type of business.
• But if the directory data are sorted as a
  sequential file, they can be in only one sorted
  order. One sequential file or records cannot be
  in sorted order by last name, first name and also
  be in sorted order by address.
• To satisfy the sort requirement, you will have
  three sequential files. One is sorted by first
  name. Another one is sorted by last name. And the
  last file is sorted by address. These three
  sequential files would contain the same data but
  be sequenced differently.

5
Processing Sequential File

• Because the records of a sequentially organized
  file must be accessed consecutively, sequential
  files are used more commonly in batch processing
  than in interactive processing.
• For example, a sequential file of customer data (
  names, addresses, credit ratings, and the like)
  could be accessed in batch mode to print mailing
  labels to advertise a special sale each customer
  record would be accessed.
• A sequential file of employee payroll data could
  be accessed batch mode to generate paychecks.

6
Advantages and Disadvantages

• The major advantage of the sequential technique
  of organizing files is the ability to access the
  next record quickly. As long as the pattern of
  access to a sequential file matches the ordering
  of records on the file, access times are very
  good. However, if a programs patterns of access
  do not match the record ordering pattern, then
  performance of the program can be terrible.
• Another advantage of sequential file organization
  is its simplicity.

7
Storing Sequential Files

• There are two basic classes of secondary storage
  devices serial-access and direct access devices.
• On a serial-access storage device, the time
  required to access two records is a function of
  the space between their locations (addresses) on
  the device. It takes longer to access them if
  they are far apart than if they are closer
  together.
• On a direct-access storage device, the time
  required to access two records is not necessarily
  dependent on their relative locations.

8
Storing Sequential Files

• Sequential files can be stored on either serial
  or direct-access devices.
• The serial-access devices are considerably less
  expensive than the direct-access devices on a
  computer system.
• For example, tape drives is less expensive than
  the disk drives.
• The direct-access storage devices on a computer
  system nearly always provide larger capacities
  and faster data access than do the serial devices
  on the machine.

9
Magnetic Tape

• The tape medium is a strip of thin, flexible
  plastic called Mylar, with a coating of
  ferric-oxide film. The tape is kept on a reel,
  which is mounted on a tape drive for read/write
  access.

• The records on the tape are stored in physically
  sequential order. The time required to access two
  records is a function of how far apart they are
  on the tape. The best case is that the second
  record would appear immediately following the
  first the worst case is that the second record
  would be at the other end of the tape.

10
Data Representation and Density

• Data are recorded digitally on the tape medium as
  magnetized spots in the ferric-oxide film
  coating.
• Positive magnetization represents a 1-bit, while
  negative magnetization represents a 0-bit, or
  vice versa.
• The magnetized areas are not randomly located on
  the medium rather they are arranged in tracks,
  which run parallel to the edge of the tape.

• There are usually nine tracks on a tape. Eight of
  them record data and the ninth records
  error-control bits.
• Density is measured in units of bits per inch
  (bpi).
• Density is a function of both the tape medium and
  the drive used to record onto the medium.

11
Parity and Error Control

• In order to interpret correctly the data that are
  stored on a tape, the encoding scheme (EBCDIC,
  ASCII, or other code) that was used when the data
  were written must be known.
• Additionally, the technique used to record
  error-control bits has to be known in order for
  that information to be used.
• One common technique for error control on
  magnetic tape is parity.
• Parity bits are written when data on the tape are
  written, then are checked when the tape is read
  to verify that an error has not occurred.

12
Parity and Error Control

• In odd parity, there must be an odd number of
  1-bits represented for each character.
• If the bits making up a characters binary
  representation already have and odd number of
  1-bits, then the parity bit (in the ninth track
  of a 9-track tape) is 0-bits if there is an even
  number of 1-bits, the parity bit for that
  character is a 1-bits, making the total number of
  1-bits for that characters representation odd.
• Similarly, parity bits on a tape with even parity
  are recorded so that there is an even number of
  1-bits represented for each character.
• For example

11110001
Odd parity
0
Parity bit
11110001
Even parity
1
13
Parity and Error Control

• This simple technique of even/odd parity can
  detect single-bit (or 3-bit or 5-bit or 7 bit)
  errors in a characters representation, but can
  not detect 2-bit, 4-bit, 6-bit, or 8-bit errors.
• Furthermore, this technique does not indicate
  which of the bits is wrong.
• Error may appear on a magnetic tape because of
  dirt deposits, bends or wrinkles, areas where the
  ferric-oxide coating is missing, worn edges,
  stretched areas, too low a level of
  magnetization, errors in recording the data in
  the first place, and so on.

14
Blocking

• Data are read from or written to a tape in groups
  of characters called blocks.
• A block is the smallest amount of data that can
  be transferred between secondary memory and
  primary memory in one access.
• A block may contain one or more records.
• A block is sometimes referred to as a physical
  record.
• Between each pair of blocks, there is a space or
  gap (sometimes called an interblock gap)

15
Tape Marks and Labels

• The beginning and end of the tape reel are each
  marked by an aluminum, reflective strip that can
  be sensed by photoelectric sensor in the tape
  drive.

• The tape mark at the beginning of the tape reel
  is known as the load point. Data can be written
  on the tape after this point until the
  end-of-tape mark is encountered.

• A piece of adhesive affixed to the reels plastic
  casing is known as external label.

• A tape may also have an internal label, which is
  recorded electronically on the tape.
• A tape with an internal label is called a labeled
  tape.

16
Performance of Sequential Files

• The performance of a sequential file is
  determined by three major factors blocking
  factor, file length, and key selection.
• Blocking Factor
• When a sequential file is stored on a
  serial-access storage device, it is generally
  desirable to make the blocks as large as
  possible. The limiting factors become main memory
  buffer space availability, operating system
  parameters (which may limit block size, and the
  local charging algorithm (which trades off main
  memory use and secondary storage accesses).
• When a sequential file is stored on a
  direct-access storage device, it is generally
  desirable to make the block size as close as
  possible to the sector size or to the track size,
  since an entire sector or track will be passed
  over to access anything on that sector or track
  anyway.

17
Performance of Sequential Files

• File Length
• File length is determined by data volumes and
  record design, which is the problem of
  determining which fields are in which records in
  which files.
• One guideline is to segment fields into records
  according to how frequently they are accessed
  together. For example consider an application
  that needs a file of employee personnel
  information. Some information about employees may
  be used more frequently than other information
  about the same employees. If so, it may be
  advantageous to split the employee data into two
  record types and to place them in separate files.

• Splitting the employee data into an active file
  and a relatively inactive file reduces the size
  of the active file and therefore reduces
  processing time on that file.

18
Performance of Sequential Files

• Selecting the Key
• The key of a sequential file determines the order
  in which its records are accessed.
• The appropriate field(s) to form the key of a
  sequential file are determined by the application
  requirement.
• For example, if the file is a report file, then
  the ordering of records is determined by the
  desired form of the output. If the file is a
  master file, the key usually is some obvious
  identifier.
• If a sequential file is not sorted on a
  particular field, records will simply appear in
  the order they were written.