BUG: ROLAP Cube Processing Fails with Big Numbers in Measures Columns

• Microsoft SQL Server OLAP Services version 7.0

SYMPTOMS

Cube processing fails when all of the following conditions apply:

• OLAP Services uses SQL Server (6.x or 7.0) as the underlying data source.



• The cube contains a measure from an integer column with large enough values that the aggregated values can exceed the valid integer range.



• The measures aggregate function is SUM.



• The Storage mode of one or more of the cube's partitions is ROLAP.

CAUSE

The SUM is performed on the column of integers, and when the value exceeds SQL Server's valid integer range (-2147483648 to +2147483647), an arithmetic overflow occurs during internal SQL Server calculations.

WORKAROUND

Before adding the integer values in this column, first convert the data type to float, or double. This can be done by changing the Source Column property of the measure to convert (float,MyIntegerColumn). The user might also want to use MOLAP or HOLAP storage options when designing the cube. After you perform one of these actions, process the cube.

STATUS

Microsoft has confirmed this to be a problem in SQL Server version 7.0.

Additional query words:

Keywords : kbSQLServ700bug
Version : winnt:7.0
Platform : winnt
Issue type : kbbug