Index Automation

Overview

Index Automation considers the following when designing indexes for the project:

• Relations between tables with relationship type set to Error or Warning
• Joins on conditional lookup fields
• Primary Key fields (On Raw Table)
• Selection Rules on the Data Warehouse
• Incremental Selection Rule on the Data Warehouse
• Partitioning fields(DW_Partitionkey, DW_TimeStamp)

Process

The Index Automation can be set on the Project level when creating a new project or through the Edit Project dialogue:

Index Automation has three associated types:

• Automatic - Always On

  When set to Automatic, the index automation updates the indexes whenever the user changes the project in a way that could trigger a new or altered index.

• Manual - Per Request

  When set to Manual, the user has to activate the Index Automation for every table where the Jet Data Manager should create indexes.

  The Index Automation is activated on the right click menu for the table (if the index-node is not present) or on the Index Node.

  

• Disabled

  When set to Disabled, the Jet Data Manager will create individual indexes during the Data Cleansing Procedures when needed.

Setting Index Generation Behavior for Individual Tables

The default settings on the project will be inherited on all tables, but you can defer from this by changing the settings on the individual tables. To do this right click the table, Advanced > Advanced Settings. The table settings window will open.

From here you can change the Index Generation Behavior for the object.

Index Creation Behavior

The indexes created by Index Automation will be named AutoIndex and postfixed with a number for uniqueness within each table.

Index Automation will try to minimize the number of needed indexes. If two lookups can utilize the same index, the Jet Data Manager will take advantage of that. When calculating which indexes are needed, the Jet Data Manager takes your manually created indexes into consideration. It will not change your manually created indexes, but it will use them instead of creating similar indexes.