Consider the following scenario:
- You have a partitioned table that contains a nonclustered index in Microsoft SQL Server 2008 R2.
- You create a clustered index that is based on the first column in the partitioned table.
- The table uses the partition key as the first column.
- You create a filter on the partition key.
- You try to create a filtered statistic for one of the columns.
- The query plan that generates a CREATE STATISTICS statement indicates that the query performed a full scan on all the partitions.
- If you create another nonclustered index, it does not take longer than expected to create the statistics for columns that are not related to the nonclustered indexes.
Microsoft has confirmed that this is a problem in the Microsoft products that are listed in the "Applies to" section.