---

Summary: Physical Data Independence

---

• Decoupling a program from the format of the data in a file
  • Use meta data:

    Data File format: +-------+-------------------------+---------+ | ID | Name | level | +-------+-------------------------+---------+ int char(30) char(4) Format description of the field: Field name | Type ---------------+---------------------- ID | integer Name | character(30) Level | character(4) Corresponding meta data: +----------------------------+-------+-------+ | ID | I | 4 | +----------------------------+-------+-------+ | Name | C | 30 | +----------------------------+-------+-------+ | Level | C | 4 | +----------------------------+-------+-------+

    ---

    ---

  • Achieveing Physical Data Independence:

    Program reads in the meta data: +----------------------------+-------+-------+ | ID | I | 4 | +----------------------------+-------+-------+ | Name | C | 30 | +----------------------------+-------+-------+ | Level | C | 4 | +----------------------------+-------+-------+ Program now knows: Each record has 3 fields Field 1 is an integer (is 4 bytes long) Field 2 is an character string and it is 30 bytes long Field 3 is an character string and it is 4 bytes long The program will read in each record according to the information in the meta data

  ---

  ---

  ---

  ---

• Advantages and disadvantages
  • Disadvantage of physical data independence:

    Program needs to use a lot of if-statements Will run slower !!! (And more complicated to code.....)

    ---

    ---

  • MAIN advantage of physical data independence:

    When the format/structure of the data record is changed, we do not need to change the program(s) !!!

    ---

    ---

    ---

  • Example:

    Consider a file of student records with the following fields:

    Field name | Type ---------------+---------------------- ID | integer Name | character(30) Level | character(4)

    The meta data file describing the data file contains:

    +----------------------------+-------+-------+ | ID | I | 4 | +----------------------------+-------+-------+ | Name | C | 30 | +----------------------------+-------+-------+ | Level | C | 4 | +----------------------------+-------+-------+

    ---

    Suppose we add a new field "address" to the Student file structure:

    Field name | Type ---------------+---------------------- ID | integer Name | character(30) Address | character(30) Level | character(4)

    Then to make the program access the new file format , all we need to do is to add another entry in the meta data:

    +----------------------------+-------+-------+ | ID | I | 4 | +----------------------------+-------+-------+ | Name | C | 30 | +----------------------------+-------+-------+ | Address | C | 30 | +----------------------------+-------+-------+ | Level | C | 4 | +----------------------------+-------+-------+

    • We do not have to change the program(s) when we change the structure of the data !!!

    ---

  • Final note:

    We DO have to convert the data file into the new format (this is an unavoidable evil..)

  ---

  ---

  ---

  ---

  ---

  ---

• Database Management Software
  • Facts:

    A data base consists of many different files. Files have different record structures The vendor of a Database Management Software system will never give you the source code And yet: The Database Management Software system must be able to read/write data files whose structures is specified by the buyer (user) of the Database Management Software system

    ---

  • Conclusion:

    A Database Management Software system must use the meta data technique !!!