About performance configurations
更新： 2009-04-30
When a client application connects to Microsoft SQL Server and performance is slower than expected, people typically examine SQL Server and determine how to optimize the query and schema for better performance. The same is true with SQL Server Analysis Services. Indexing your ID-based and label-based tables is one possibility. Individual uses will determine the best solution for your project.
The following are indexing recommendations the might help you improve performance.
Considerations for indexing
During synchronization in Step 2: Synchronizing the staging database, the operation will drop the tables so that all indexes must be recreated.
When you convert ID-based tables to label-based tables during Step 3: Populating the staging database, the stored procedure will drop the tables and the indexes must be recreated.
Indexes for ID-based tables
Dimension
BizSystemFlag, MemberId, and Label columns (Composite Index)
BizSystemFlag, Label and MemberId (Composite Index)
BizSystemFlag and MemberId (Composite Index)
MemberId and Label (Composite Index)
All Columns that have FK
Hierarchy
BizSystemFlag and MemberId (Composite Index)
ParentMemberId
BizSystemFlag
Measure group and annotations
All Columns that have FK
BizSystemFlag
Indexes for label-based tables
When converting label-based tables back to ID-based tables during Step 3: Populating the staging database, we recommend that you do not have any indexes on target ID tables.
Dimension Label table
All Label columns that have FK
Hierarchy label-based table
MemberId_Label
ParentMemberId_Label
Measure group and annotations
- All Label Columns that have FK
Download this book
This topic is included in the following downloadable book for easier reading and printing:
See the full list of available books at Downloadable content for PerformancePoint Planning Server.