Data integrity 

..to previous page[5] 


Integrity refers to data that are;
* accurate,
* reliable and
* consistent.
In order to maintain data integrity, data needs to be accurately collected, entered into the database, and managed by the database.
Two types of systems allow for data entry; batch and on-line.
  • Batch allows for data to be saved and put through as a "bundle". For example
    • Student enrollment systems where paper forms are filled in by students and sent to the institute for processing
    • Banking systems where the days transactions and saved separately and merged with the main files "overnight".
  • on-line updates records immediately after they are entered. For example;
    • Flight reservation systems
    • Hotel room bookings systems

Source checking
  • Source checking can occur in many places;
    • when the paper or screen form is being filled out.
    • when the data is being entered into the database (maybe via screen or communications device like a modem)
  • To assist in source checking;
    • forms may be identified with a code and often numbered (eg statements, cheques)
    • processing of documents is logged - date in, date out either individually or as batches
    • records are given unique identifiers, and may be date/time stamped.
    • the number accepted is matched against number rejected and number input
  • Source checking may also include;
  • Range
    • Is the content within a particular range, or list (eg month is 1-12, sex is F or M)
    • Should the content be a member of a particular list, either in a file or in a table. (e.g. valid department numbers )
  • Optional/manditory
    • Is field optional or essential (manditory)
    • For example: Surname
  • Reasonableness
    a minimum and/or a maximum range is set.
    • For example:a person attending the Polytechnic will be between 5 and  90, withdrawals from an ATM may be up $200 a day.
  • Pattern
    • Does a code have a particular pattern or algorithm associated with it
    • For example: phone number, credit card number, check digit, bar code
  • Consistency
    • For example : if you were entering a telephone number for a person living in Taradale the telephone number would start with (06)844.., Tamatea area (06)843..
  • Dependency
    • Does the value in this field relate to any other field
    • For example: If a wage is calculated has the hourly rate been entered.
  • Type
    • data consists of only certain types of characters.
    • For example: IRD numbers are 8 numeric characters separated by slashes, a persons name is alphabetic characters.
Audited formulae   Check totals

If a form contains a formula for example wages = hours worked * hourly rate, this formula should be checked by using simple numbers. Since the computer stores numbers in different ways, a formula needs to be checked to ensure it will cope with possible answers.

  • For example: a formula with a data type of integer will have difficulty dealing with currency (2 decimal places).

When entering data check totals can be used

  • For example
    The number of records entered equals the number of records counted (checks for double entry)
    The total value of the entered transactions equals the amount calculated from the source documents (checks that numbers have been entered correctly i.e $100 not $1000!).

[Rev 27/06/00] 1996 © 1996-2000 V/2-Com (Verhaart), P O Box 8415, Havelock North, New Zealand.