Data Quality: The Accuracy Dimension (The Morgan Kaufmann Series in Data Management Systems)

B.5 Table Definitions

• Name

• Descriptive name

• Business meaning

• Columns

  • Name

  • Longer descriptive name

  • Business definition

  • Confidence indicators

    • Trusted

    • Susceptibility to decay

    • Enforcement processes

  • Domain names if inherited

  • Data type

    • Data type discovered

  • Length boundaries

    • Maximum length discovered

    • Minimum length discovered

    • Length distributions discovered

  • Numeric precision

    • Maximum precision discovered

  • Value properties

    • Discrete value list

      • Values discovered

      • Inaccurate values

      • Values not used

      • Value frequency pair list discovered

      • Encoded value meanings

    • Range of values permitted

      • Range of values discovered

      • Skip-over rule

        • Skip-over rule violations

    • Character patterns required

      • Patterns discovered

    • Character set

    • Character exclusions

    • Text field restrictions

      • Keywords discovered

      • Text constructs discovered (embedded blanks, special characters)

      • Upper/lowercase conventions discovered

      • Leading/trailing blanks discovered

• Property rules

  • Unique rule

    • Uniqueness percentage discovered

  • Consecutive rule

    • Consecutive rule discovered

  • Null rule

    • Null indications

      • Null indications discovered

      • Blank or zero rule

  • Inconsistency points

    • Date of change

    • Description of change

• Functional dependencies

  • LHS columns

  • RHS columns

  • Type

    • Primary key

      • Token

      • Natural

      • Denormalized key

      • Derived column

        • Rule or formula

    • Discovered percentage true

      • Violation values