New Column Mapping
Use this dialog to define or edit a mapping for a transactional column in a joined table, that is, a One to Many relation.
For example, you could create a mapping that aggregates by sum to calculate the total amount of product sold to a given customer grouped by product ID.
To define a mapping, specify the following information:
- Value Column: Select the column in the transactional table; the values in this column will be aggregated.
- Mapping Name: Specify name of the value column. If the Automatic box is checked, the wizard will generate a mapping name based on the column name. The default is to generate names automatically. To specify a mapping name, uncheck the box, and type the name.
- Mining Type: Specify the mining type of the column. Select a mining type from the pulldown. Not all columns can be Value Column. If you see
Not Allowed
, you must select a different Value Column.
- Group By: Specify how the data should be grouped; You can group by
<Case>
or by a column of the table being joined.
- Aggregation: Select the aggregation for the transactional column. The choices are
SUM
,
AVG
(average), MIN
(minimum), MAX
(maximum), COUNT(DISTINCT)
(number of distinct values), COUNT
, and MODE
.
- Unique Prefix: This read-only field contains the unique prefix that is automatically generated by the wizard.
- Sparsity: The default is to check Data is Sparse; if the data is not sparse, uncheck the box. Note that not all activities support mapping to sparse data; for example, attribute intelligence models cannot be built using sparse data. If the activity that you are defining does not support mapping to sparse data, you will get a message after you click OK.
Click OK. This wizard validates the definition. If it finds a problem, it generates a message. When there are no problems, you return to the Edit Relationship dialog.
Copyright © 2006, 2008, Oracle. All rights
reserved.