How ETL process works in SAS IFRS9

[Musheer1]

Hi everyone,

I would like to better understand how the ETL process works in SAS IFRS9.

Specifically, I have a few questions:

1. Does the ETL process load data into the IFRS9 Input Area, or directly into the Landing Area?
2. What is the difference between the Input Area and the Landing Area in terms of their roles in the ETL pipeline?
3. At which stage are the files moved from Input Area to Landing Area?
4. How does this impact downstream jobs such as Data Preparation and Credit Risk Analysis?

Additionally, if files are uploaded through a stored process:
5. How can we verify that the uploaded files are correctly reflected in the backend (e.g., properly loaded and available for processing in the ETL flow)?

I would appreciate a clear explanation of the end-to-end data flow and validation steps.

Thanks in advance!

[cyezer · Accepted solution]

Hello, thank you for the great set of questions—these go right to the heart of how the SAS Solution for IFRS 9 on SAS9.4 data pipeline operates. Based on the solution architecture (see User Guide, Chapter 2: Architecture and Workflow), here is a summary of that explanation (however I highly suggest looking at the chapter as there is a nifty data flow architecture visual design).:

1. Does ETL load into the Input Area or Landing Area?
In SAS IFRS 9, data is ultimately loaded into the Landing Area, which serves as the formal entry point into the solution. Any upstream ingestion (manual upload, stored process, or external ETL) should result in datasets being placed there.  The out-of-the-box SAS9.4 software does not orchestrate ETL processes from your company's proprietary databases into the Landing area; this would be configured during setup from an upstream process.  For more information on registering data into the Landing area, please see the SAS Risk Stratum Cookbook: Configuring data: Registering Input data.

2. Input Area vs. Landing Area

• Landing Area: The official staging location used by IFRS 9 processing. Data here is in a structured format ready for ingestion.
• Input Area (informal term): Sometimes used to describe pre-staging locations (e.g., uploaded files or external feeds before they are standardized). It is not a formal processing layer in IFRS 9.

In practice, only the Landing Area participates directly in the IFRS 9 Solution ETL pipeline.

3. When are files moved to the Landing Area?
Files should be placed in the Landing Area before running any IFRS 9 cycle.
The key transition is not a physical “move” within IFRS 9, but rather the registration step during Data Preparation, where data in the Landing Area becomes recognized by the application.

4. Impact on downstream jobs (Data Prep, Credit Risk Analysis)

• Data Preparation reads data from the Landing Area, applies transformations/enrichment configured by you and stored in your configuration tables, and registers the output tables into the IFRS 9 data repository as Analysis Data.  Analysis data is stored in a central repository and is accessible by all SAS Risk solutions installed.  Each solution accesses the data using a view, rather than copying the physical data into the solution.
• All downstream processes (Data Quality, Adjustments, Credit Risk Analysis, reporting) operate only on this registered data.  Registered data cannot be modified or overwritten.  
• If data is missing or not properly staged/registered, downstream jobs may fail or run with incomplete inputs.

5. Validating stored process uploads
To confirm that uploaded files are correctly available for processing:

• Verify datasets exist in the Landing Area library/folder
• Run Data Preparation and confirm successful completion
• Check that Analysis Data objects are created (correct base date, metadata)
• Review job logs for errors or warnings
• Optionally, inspect linked relationships (Analysis Data ↔ Data Definitions ↔ Cycle) to confirm proper registration

End-to-end flow (simplified):
Upload / ingestion → Landing Area → Data Preparation (register + enrich) → IFRS 9 Repository (Analysis Data) → Downstream processing (risk, reporting, disclosure)

Hopefully this explanation and the documentation links help provide clarity to your question.  If you have follow-up questions, could you provide more information regarding the ETL flow mentioned?  Thank you!