MDXINSERTBUFFERAGGMETHOD

When you perform an MDX Insert, Essbase creates an output buffer in memory, which accumulates with values until the Insert is completed.

The MDXINSERTBUFFERAGGMETHOD configuration setting enables you to control resolution of conflicts in the buffer. The method you use can impact the results of the Insert. By default, duplicate cell values are summed together, but you can configure Essbase to use the value of the cell that was loaded last.

Syntax

MDXINSERTBUFFERAGGMETHOD ADD | LAST

LAST—If, during an MDX Insert operation, a value needs to be written to an output buffer location that already contains a value, the latest value overwrites the older value.
ADD—If, during an MDX Insert operation, a value needs to be written to an output buffer location that already contains a value, the latest value is summed with the older value. This is the default behavior.

Description

During execution of an MDX Insert query, an output buffer is created in memory which accumulates with values until the query is completed. MDXINSERTBUFFERAGGMETHOD enables you to define the method with which values are aggregated in the output buffer. The method that you use can have an effect on the data results of the MDX Insert operation.

Assume that in an MDX Insert query, two source tuples are mapped to a single target tuple, as shown:

INSERT
([Payroll], [Jan]) TO ([Revised Payroll], [Jan])
([Payroll], [Feb]) TO ([Revised Payroll], [Jan])
...

Assume that the value of ([Payroll], [Jan]) is 100, and the value of ([Payroll], [Feb]) is 200.

Using the default buffer aggregation behavior (ADD),

MDXINSERTBUFFERAGGMETHOD ADD

The value for ([Payroll], [Jan]) is written to the output buffer for ([Revised Payroll], [Jan]), making its value 100.
The value for ([Payroll], [Feb]) is added to the same output buffer for ([Revised Payroll], [Jan]), increasing its value to 300.

If you change the buffer aggregation behavior to LAST,

MDXINSERTBUFFERAGGMETHOD LAST

The value for ([Payroll], [Jan]) is written to the output buffer for ([Revised Payroll], [Jan]), making its value 100.
The value for ([Payroll], [Feb]) is written to the same output buffer for ([Revised Payroll], [Jan]), overwriting the previous value, and changing it to 200.