The ordinary Tableau FILTERS are mutually autonomous. Each filter reads all lines and generates its own outcome from the source information. There may, however, be situations where only the results returned by the first filter can be processed by the second filter. In such a case, the second filter is referred to as dependent filters because the data passing through the context filter is processed only. Background Filters are used for two primary reasons.

• Improves efficiency–The queries can be slow if you set many filters or have a big information source. To enhance efficiency, you can set one or more context filters.
• Create a number or top N filter − A context filter can only include interest data and then set a numerical filter or a top N filter.

Any other filters you set are described as dependent filters because only the information PASSED through the context filter is processed. Filters used to enhance efficiency are often used in context. If the context filter does not decrease RECORD numbers by 10% or more, the dashboard can effectively be slowed.

The user does not change the context filter many times–when the filter is changed the database has to be re-published and the temporary table re-written, slowing its efficiency.

When you set a context dimension, Tableau creates a temporary table to be reloaded when the view has been launched. The temporary table generated is available in the  Access Table format for Excel, Access, and text information sources. You must have APPROVAL to build a Temporary Table on your server for SQL Server, My SQL and Oracle information sources. Temporary tables or context filters are not created in multidimensional information sources or cubes, and only which filters are separate and dependent.