Reasons Not to Use Junk Dimensions

Junk dimensions can be a pain in the neck. Often people will say that it is better just to add extra columns to the fact table, especially if the junk dimension represented only 'Yes' or 'No'. Here are some reasons why:

• If you ever want to update junk dimensions, it can involve a change in the meaning of existing members (cells) in the junk dimension.
• It is hard to maintain a surrogate key for a junk dimension
• If you ever want to move an attribute (column) of a junk dimension into a fully fledged dimension, this will break all the reports that previously used that key from the junk dimension.

Nevertheless, junk dimensions do have these advantages:

• Junk dimensions reduce the number of dimensions within a cube, making it easier to browse
• Reducing the number of dimensions means the cube performs better during querying