ERP Automation

 Automation Approach for ERP

Agenda

• ERP Automation Approach
• Selenium Automation Framework Overview
• ERP Automated Functional Tet Case Demo
• ERP Automated Pipeline Trigger Demo
• ERP Automated Integration Test Case Demo

Automation Preconditions:

Document Gathering: 

• Finalized design diagram
• Requirement document 
• Detailed test scenarios

Access Required:

Access required as per the application requirement (ALM, ALM Excel Add-Ins, Bitbucket, Terminal Server, Oracle Fusion Application, Azure Data Storage File Location, User Credentials access, Automation Environment)

Identify Technology and Tools:

Feasibility analysis of test scenarios
Oracle Fusion application for UI Validation

Categorization and Prioritization:
Prioritize test cases based on biggest business impact and categorized test cases based on functionalities (Functional flows, Integration Flows, Batch runs)

Automation Framework and Scripting approach:
New keyword functions and locators will be added to support new functionality

Existing Selenium for web-based application

Automation Design Appraoch for SIT1: 

Part 1 of Automation Script:

All Oracle Fusion Functional Flow for all scenarios

Part 2 of Automation Script:

Trigger Azure Integration Pipelines
FUSION_EXTRACT_RUN, FUSION_EXTRACT_DECRYPT, FUSION_EXTRACT_VIRUSFREE, FUSION_EXTRACT_OUTBOUND

Part 3 of Automation Script
Verify the content of generated file for all Integrations for all scenarios

Ways to trigger pipeline:

Test Suite to trigger pipelines will have test cases for both of the following, giving us the flexibility to trigger pipelines as needed:

Triggering pipeline for all integrations at once
Trigger pipeline for each integration separately

Automation - Total Coverage, Planned Test Scripts for E2E coverage, Total Test Scripts Delivered

E2E Scenario: New Hire of Full Time Regular Union Employee

1.Functional Flow in Oracle Fusion

• Recruiter assigns candidates to the requisition
• Recruiter creates, extends and accepts offer
• Recruiter accepts offer on behalf ofa candidate
• Recruiter moves candidate to HR
• Convert the pending worker to employee
• Probation calculation extension
• Compensation Admin Run GSP Extension
• Compensation Admin Run gsp pROCESS
• HRDA validate absence plan enrolment and accruals

2.Trigger Pipeline

3.Verify the content in the files

E2E Scenario: Terminate a Full Time Union Temporary Employee

1.Functional Flow in Oracle Fusion

• HRDA initiates Termination
• Notification to Manager
• Compensation Admin - Assign Termination Notice - Temp EE ICP
• HRDA validate absence plan enrolment and accruals

2.Trigger Pipeline

3.Verify the content in the files

CDS Automation Approach:

SCENARIO

Traditional way of comparing CDS files (csv to Parquet) is time consuming taking ~30-60 mins per file comparison, error-prone, and couldn't scale with growing data volumes

TASK

To automate this "Sort-and-Compare" workflow within our Selenium framework to handle multiple large different files (csv and parquet) without manual effort.

CHALLENGE

• CSV treats everything as text; Parquet uses binary types.
• Hardcoding column names means tests break when the schema changes.
• Large files (2GB+) crash the JVM due to Heap limits.
• Comparing 100+  columns across millions of rows takes minutes.

SOLUTION: DuckDB Engine

What is it? An in-process SQL OLAP engine that requires no server installation.
Why DuckDB? 

• Treats .csv and .parquet like database tables instantly.
• Uses the EXCEPT operator to find differences in seconds
• Extremely low memory footprint; it "streams" data rather than loading it all.

Workflow Implemented for CDS Automation:

• Read Parquet schema without hardcoding column names
• Normalize the values automatically vis SQL to prevent "false failures" from white spaces or casing.
• Eliminate nose by re-structuring the files in memory to have exact same column order and then sort every row by every column
• One-Line comparison by executing a single query to find records in the CSV missing from Parquet.
• Actionable Reproting: If a mismatch is found, automatically trigger COPY TO 'mismatch_report.csv' for analysis.