Data Manager allows you to upload and manage all the data in Analytics.
...
Raw external data is loaded into a Data Feed from an external source:
Via the Server REST API using:
IntegrationManager – a Pricefx application
Customer's own software
By an Excel file upload (directly via user Interface in Data Manager).
Data Feed data can be unstructured (all the columns are strings) and can include unwanted or duplicate data (there are no key columns).
Data Load Flush – With the Flush Data Load, the data is loaded into a Data Source. You can use a calculation logic to structure, clean up and validate the data.
A Data Source always has a key – it cannot contain duplicate data. Also, all columns have strict data types (like Number, String, Date, etc.).Data Load Refresh joins together different Data Sources into a Datamart, based on one or more keys.
Additional values can be calculated in the Datamart using expressions at query time or via a special Calculation Data Load using a logic.
When Publishing Data Load is run, it exposes the ‘refresh’ data to Datamart queries.
Note: The published ‘refresh’ data table is designed assumed to be read-only, taking advantage of column-store performance benefits, while the source data continues to be stored in regular row-oriented tablespublished until the first run of the Publishing Data Load.
Drawio | ||||||||||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
...
The calculation will not have a Data Source as a source. The "source" DatamartRowSet will be the Datamart itself.
When you want to source data from a Data Source, to load them in the Datamart you will need to both query the Data Source and to use the "target" DatamartRowSet's addRow(map) method.
This way a Datamart can be entirely populated by a logic (sourcing the required data using api.finds, Analytics queries etc.).
But you cannot use the 'free' joining of source Data Sources and also the incremental load mechanism (i.e., you need your logic to do all that, if required).
...