Predictors in Adaptive Models
Can we add predictors from other tables(not from customer table) into the adaptive model? We have multiple tables joined with customer table in our application. The customer class is defined with "has a" relation associated to the customer class. Please let me know if there is any limitation on this. Pega Marketing 7.3.1
As long as the reference is visible and is a scalar (flat, not an array reference), it can be referenced by the model.
Customer.Address.Postcode [Works]
Customer.Products(5).Value [Does not work]
To make it work you would need to flatten in to Customer.Product_5.Value, where Product_X is a known and existing page to the system.
Hi Saleem_A,
Thanks for sharing your response. It is saving if we use the page properties in adaptive models but it is not able add the pxADMInputs to the decision results record. Is it a product defect PM 7.3.1? Lets think in other way: This is my analysis. The predictors are picked from the customer table columns right. If we refer .Address.Postcode, it means there should be a dedicated column to this table.
If we do not have dedicated column in the customer table, then how does pega picks the column from the other table? The page created(Address page in customer class) in the customer class is a clipboard relationship with the customer class but the table relationship. Correct me if I am wrong. Thanks in advance.
Regards,
Nizam
In DMSample we're using constructs that index into a page list and take some property - no need to flatten.
BTW I noticed related posts around the usage of multiple, linked, data sources as predictors, e.g. https://pdn.pega.com/community/product-support/question/read-adaptive-p….
it is not about which and how many data base tables, it is all about different pages in clipboard in runtime when your strategy is running.
in Adaptive Model rule you can refer properties from apply-to-class (typically Customer class) or/and you can refer some properties from the pages which is defined in the "page and class" tab, as predictor.
now whether and which properties will be a good predictor for your business use case that your data analyst can suggest.
Regards
Shankha