Best Practices

Use members from the dimension main hierarchy for improved report performance. The main hierarchy contains data and rollup members. The data members store data. You should use data members directly in a query to improve report performance. Data for the rollup members is calculated using the value of the children and the rollup operator.

It is best to use rollup operators on the Account dimension only. Use the default rollup operator ‘+’ for all other dimensions.

Avoid using Alternate hierarchies unless they are necessary. Alternate hierarchies contain reference members. Report performance may be decreased when you use reference members in reports.

Use an Attribute hierarchy to create an alternate hierarchy.

Avoid adding too many members under one parent (flat hierarchy), for example, rolling up days information directly to year level.

Keep Segment Hierarchies simple.

Consider loading unused/discontinued members to one member under the hierarchy.

Don’t forget to set a Member Code for each of your Segment Hierarchies.

Load all segment hierarchy data prior to creating budget hierarchies and entities.

Entities may be added to a hierarchy at any time.

Avoid adding too many members under one parent (flat hierarchy).

Creating a very large Budget hierarchy can impact performance. A large hierarchy is 5k to 10k members. A very large hierarchy is over 10k members. This applies to all hierarchies.

Entities may only be deleted when data is not associated.

Avoid having too many levels.

Don’t forget to set a Member Code for your Budget Hierarchy.

Planful delivers system-define attributes.

For Financial Segments, you may have up to 25 attributes per dimension.

For Sales, you may have up to 10 attributes per dimension.

You can build two types of Attribute hierarchies; Rollup Based on Attributes or Rollup Created Manually. Use Rollup Based on Attributes when possible. Manual rollups are used to build the hierarchy similar to the chart of account hierarchy.

Avoid having too many members under one parent. For example, one parent with 300 children is a good candidate for revision.

Your attribute hierarchy should contain no more than 5 to 7 levels.

The List type attribute offers the best performance over any other attribute type.

When you have TEXT, NUMERIC, or DATE type attributes, add them as member properties to the cube rather than to the attribute itself.

Attributes increase the overall cube space so use them accordingly. Try not to exceed 5 attributes on a single dimension. Try limiting the number of attributes (8 to 10).

Not all attributes are required for reporting or slice or dice.

Dimension security is available for attributes, but data level security leads to performance implications.

Attributes perform better versus member properties in the reporting cube. It’s best to configure less than 3 attributes as member properties.

The maximum number of attributes that can be mapped to a cube is 5.