| Feature | Azure Data Factory | SSIS (On-Prem) | |---|---|---| | Execution | Serverless (pay per run) | Requires dedicated server | | Scale | Auto-scales thousands of activities | Manual scale (more workers) | | Maintenance | Microsoft handles patches | DBA team required | | Hybrid Access | Self-Hosted IR | Gateway or VPN | | Cost Model | Consumption (DIU hours, pipeline activity) | Licensing + infrastructure | | Learning Curve | Low (UI based) | High (complex components) |
Lift & Shift Strategy: Use Azure-SSIS Integration Runtime. You can redeploy existing SSIS packages (.dtsx) into ADF without rewriting.
Following the Javatpoint teaching methodology, let's build a practical ETL pipeline using the Azure Portal. Our goal: Copy data from a public blob storage (Source) to an Azure SQL Database (Sink). javatpoint azure data factory
| Aspect | Javatpoint | Microsoft Docs | | :--- | :--- | :--- | | Tone | Simplified, beginner-friendly, exam-focused (e.g., DP-203, DP-900) | Comprehensive, technical, reference-oriented | | Examples | Small, isolated examples (e.g., copy from Blob to Blob) | Real-world enterprise patterns including error handling | | Best For | Students, fresh graduates, career changers | Professional developers, data engineers, architects | | Depth | Covers 70-80% core concepts | 100% coverage including edge cases |
In an era where Medium articles are locked behind $5/month subscriptions and video courses cost $200, Javatpoint remains completely free. No credits, no “start your 7-day trial.” For students in developing countries or self-funded learners, this is not a minor advantage—it’s a lifeline. | Feature | Azure Data Factory | SSIS
| Transformation | Purpose |
|---|---|
| Source | Reads from a dataset (JSON, Parquet, CSV). |
| Filter | Removes rows based on condition (e.g., Price > 100). |
| Derived Column | Creates new columns or modifies existing ones (e.g., Total = Price * Quantity). |
| Aggregate | Group by and compute sum, avg, min, max. |
| Join | Combines two streams (Inner, Left Outer, Full Outer). |
| Sink | Writes transformed data to destination (Delta Lake, SQL, ADLS). |
Input: Raw clickstream logs (JSON). Flow: Source → Filter (remove bots) → Derived Column (extract date from timestamp) → Aggregate (count clicks per product) → Sink (Azure Synapse Analytics). Following the Javatpoint teaching methodology, let's build a
Why use Data Flows? They allow non-programmers (BI analysts) to perform complex ETL without coding Spark.