Limitations when using Context Filters

Under certain conditions, you might encounter unexpected results when using context filters. For example, measures in the analysis might appear as "--" in IBM® Cognos® Analysis Studio.

This issue can occur under the following conditions:

  • In IBM Cognos Framework Manager, the Aggregation rule of the measure is set to Last.
  • The dimension is expanded in the analysis.
  • The dimension being expanded is inserted into the context filter area.

In general, avoid creating a context filter that uses items from the same hierarchy as an item on one of the crosstab edges; instead, put the item directly on the edge. For example, instead of placing all years on the crosstab edge and placing 2005 in the context filter, simply put 2005 on the crosstab edge.

If you insert members from the same hierarchy on both edges of the crosstab, you may encounter unexpected results. For example, an analysis that uses members from Years in the rows and Quarters in the columns is difficult to read because the useful numbers are distributed over a large area of mostly blank cells.

Here are some further considerations when using context filters:

Calculated Measures Are Not Suppressed by the Context Filter

You are working with dimensionally modeled relational data and you have multiple measures on the report, one of which is a calculated measure (that is, the Regular Aggregate property is set to calculated in the model). You apply a context filter using items from the same hierarchy as an item on one of the crosstab edges, but values that do not pass the context filter are not removed from the result, even if suppression is applied.

An example of a context filter in an analysis. There is a problem if you apply a context filter using items from the same hierarchy as an item on one of the crosstab edge. Values that do not pass the context filter are not removed from the result, even if suppression is applied.

To avoid this, you can drag the desired members from the Insertable Objects pane directly onto the row or column edge instead of a using a context filter.

Note: To insert a single member, drag the 2005 item to the rows in the crosstab while keeping the right mouse button pressed, and select Insert without details.
An example of a context filter in an analysis. If there is a problem because you apply a context filter using items from the same hierarchy as an item on one of the crosstab edges, you can drag the desired members from the Insertable Objects pane directly onto the row or column edge instead of a using a context filter.

General Suppression is Slow on Large Cubes

You are using a general suppression option, such Suppress Rows and Columns, Suppress Rows Only, or Suppress Columns Only, in conjunction with a context filter.

An example showing how using suppression in conjunction with a context filter can slow down the reporting time.

If the data set is large, this can take a long time because all of the data is retrieved and then filtered. On large cubes, other means should be used to focus the report:

  • If the desired data is from a hierarchy that appears on the row or column edge, do not use a context filter. Instead, you can drag the desired members from the Insertable Objects pane directly onto the row or column edge.
    Note: To insert a single member, drag the item to the crosstab while keeping the right mouse button pressed, and select Insert without details.
    Another example of how you can drag the desired members from the Insertable Objects pane directly onto the row or column edge instead of a using a context filter.
  • If the desired data is from a hierarchy that does not appear on the edge, use a context filter in conjunction with other means to limit the data, such as a top or bottom filter, the Suppress Rows of Selection or Suppress Columns of Selection suppression option, or filter the item to limit results, using criteria such as where X is greater than Y, or where X is not null.