Wednesday 28 March 2012

Aggregation Type

Use

You use the aggregation type to control how a key figure or data field is updated to the InfoProvider.

Features

For InfoCubes:
Depending on the aggregation type you specified in key figure maintenance for this key figure, you have the options Summation, or Maximum or Minimum. If you choose one of these options, new values are updated to the InfoCube.

The aggregation type (summation, minimum & maximum) specifies how key figures are updated if the primary keys are the same. For new values, either the total, the minimum, or the maximum for these values is formed.
For InfoObjects:
Only the Overwrite option is available. With this option, new values are updated to the InfoObject.
For DataStore Objects:
Depending on the type of data and the DataSource, you have the options Summation, Minimum, Maximum or Overwrite. When you choose one of these options, new values are updated to the DataStore object.

For numerical data fields, the system uses characteristic 0RECORDMODE to propose an update type. If only the after-image is delivered, the system proposes Overwrite

However, it may be useful to change this: For example, the counter data field “# Changes” is filled with a constant 1, but still has to be updated (using addition), even though an after-image only is delivered.

The characteristic 0RECORDMODE is used to pass DataSource indicators (from SAP systems) to the update.

If you are not loading delta requests to the DataStore object, or are only loading from file DataSources, you do not need the characteristic 0RECORDMODE.

Summation:
Summation is possible if the DataSource is enabled for an additive delta. Summation is not supported for data types CHAR, DAT, TIMS, CUKY or UNIT.

Overwrite:
Overwrite is possible if the DataSource is delta enabled.

When the system updates data, it does so in the chronological order of the data packages and requests. It is your responsibility to ensure the logical order of the update. This means, for example, that orders must be requested before deliveries, otherwise incorrect results may be produced when you overwrite the data. When you update, requests have to be serialized.

Example

You are loading data to a DataStore object. In this example, the order quantity changes after the data is loaded into the BI system. With the second load process, the data is overwritten because it has the same primary key.

First Load Process
Document No.
Document Item
Order Quantity
Unit of Measure
100001
10
200
Pieces
100001
20
150
Pieces
100002
10
250
kg

Second Load Process
Document No.
Document Item
Order Quantity
Unit of Measure
100001
10
180
Pieces
100001
20
165
Pieces

No comments:

Post a Comment