Parametric Groups

Use the Parametric Group tool to create columns whose values are based on queries. The query results are saved in ‘bins’, which are essentially groups of records that share common criteria as specified in the source query. Groups are not necessarily mutually exclusive, which means that individual records may exist in more than one bin.

For example, create columns at the customer level that summarize shopping behavior based on information in the foreign ‘Order’ table. In this example the bins could be things like “Shops Online”, “Shops instore”, or “Catalogue Shopper”. Customers who make purchases in more than one way will exist in multiple bins.

Procedure

1. Drag the Parametric tool to the Workspace or select it from the New drop-down on the ribbon bar.
2. Enter a Display Name. If you save the tool as a template, this name will be used to display it in the Templates tab. The Display Name is also used in the Caption area if you insert the tool into a document.
3. If you want to make the parametric column permanent, so that it is available to view and select in the Data Explorer, expand the optional settings:

In the Table Column Name field, enter the name you want to use to display the column in the Data Explorer tab.

4. Populate the remaining Settings fields:
   • Target Table: Drag a table from the Data Explorer that will provide the final resolution of the parametric group. Note that the target table must be linked to any tables on which the segment queries are based.
   • List of Groups: This where you name and create the groups that will constitute your parametric. Click the Add button and then overwrite the ‘Group Name’ with the required name.

With a group selected, drag segments from the Data Explorer into the area on the right. You can add multiple segment parameters using the available join operators. Where your segment query contains elements from different tables, then the order that you add the elements is important. The segment resolution will be set by the first column that you drag in. The entire segment will be resolved at this level, before being resolved to the specified Target Table (if it is different).

For example, in the following segment, the first query element is Gender = Female from the Customer table. The second query element is Product Code = 'LIOG3' from the Order table. The segment count will therefore be female customers who have ordered the product 'LIOG3' i.e. 14,156

In this example, the query elements are reversed with the resolution being set by Product Code = 'LIOG3'. The segment is therefore a count of orders of the product 'LIOG3' made by females. Where a female had ordered 'LIOG3' more than once then each order would be counted separately, which explains why the segment count is higher than the first example i.e. 17,131.

The group display on the left will display two counts, a Segment count and a Group count. These counts will be different where the segment resolution is different to the group resolution as set by the Target Table field.

5. Click Test Groups to confirm that the parametric is valid.
6. Click Run Processes to create the parametric remap.

• Data Engineering
  • Data Engineering Tools Overview
  • Aggregate Functions
  • Aggregate
  • Auto Numeric Band
  • Copy Values To
  • Days Between
  • Decode
  • Engine Cleanup
  • Expression
  • Engine Expressions
  • First or Last
  • Make Permanent
  • Numeric Band
  • Parametric Groups
  • Percentile
  • Prioritize List
  • Quantile Band
  • Rank By X