Documentation
Start typing to search…

Getting Started with Conversor DDF

Passo a passo de como iniciar sua jornada na Dadosfera

Getting Started with Conversor DDF

Follow these 5 steps to convert your Tableau Prep flow or DataStage job to validated SQL.

Step 1: Upload Your Flow or Job

Upload your Tableau Prep flow or IBM DataStage job to the system.

  • Click the Upload button on the Workflows page
  • Select your file (maximum 100MB)
    • Tableau Prep: .tfl or .zip files
    • DataStage: .dsx files
  • The system will parse and analyze your file
  • You'll see the workflow details and available outputs

What happens next: The system extracts your workflow structure, identifies all nodes, and prepares conversion options.

Step 2: Create a Conversion Session

Select an output node and initialize the conversion for that specific target.

  • Choose which output node to convert
  • Configure input mappings if needed
    • Excel table references (Tableau Prep)
    • DataStage datasets
  • Review the stage count and execution order
  • Click "Start Session" to begin conversion

What happens next: The system creates a session targeting your selected output and prepares to generate SQL stage-by-stage.

Pro Tip: If your workflow has multiple outputs, create a separate session for each one.

Step 3: Work Through Stages Sequentially

The system guides you through stages, automatically blocking downstream work until upstream dependencies are complete.

For each stage:

  1. Start with the first stage (usually your input)
  2. Generate the SQL for each stage (AI-powered or manual)
  3. Use the 6 validation panels to review your work:
    • Generated SQL
    • Node Configuration
    • Intermediate Query
    • Schema
    • Feedback History
    • LLM Conversation (future)
  4. Approve, edit, or request changes
  5. Move to next stage (blocked stages unlock automatically)

Validation Workflow:

For each stage:
├─ Review Generated SQL
├─ Check Node Configuration
├─ Test Intermediate Query
├─ Verify Schema
├─ Check Feedback History
└─ Approve or Request Changes

Key Concept: Downstream stages remain locked until you approve upstream stages. This ensures data flow consistency.

Step 4: Use Batch Operations (Optional)

Speed up conversion with batch generate and reset operations.

Batch Generate:

  • Generate multiple eligible stages at once
  • Only stages with met dependencies will actually generate
  • Saves time on large workflows

Batch Reset:

  • Reset multiple stages back to pending status
  • Useful if you need to revise multiple stages
  • Dependency rules apply automatically

Pro Tip: Use batch operations after making SQL edits to regenerate dependent stages all at once.

Step 5: Export Your SQL

Once all stages are validated, export your complete SQL script for use in your data warehouse.

  • Review the final stage with all upstream CTEs
  • Export the complete SQL script
  • Use in your data warehouse
  • Version control your SQL alongside your workflows

What you'll get: A single SQL file containing all Common Table Expressions (CTEs) and the final SELECT statement, ready for production use.

Key Concepts

📊 Workflows

Your uploaded Tableau flows or DataStage jobs. Each workflow can have multiple versions that you can update without losing conversion history.

🎯 Sessions

A conversion session targets one specific output node. If your workflow has 3 output nodes, create 3 sessions to generate SQL for each output.

📋 Stages

Each node in your flow becomes a stage. You validate and approve each stage's SQL individually. Stages are processed sequentially with dependency management.

🔗 Dependencies

Downstream stages are automatically blocked until their upstream dependencies are generated and validated. This prevents invalid queries and ensures data consistency.

Pro Tips

💬 Use the Intermediate Query Panel

The intermediate query shows your complete SQL up to the current stage. Test it in your database to validate data flow and catch issues early.

🔄 Use Batch Operations for Efficiency

Generate multiple stages at once, but remember only eligible stages (with dependencies met) will actually generate. Non-eligible stages are automatically skipped.

✏️ Don't Hesitate to Edit the SQL

The AI-generated SQL is a starting point. Edit freely to match your requirements. Your edits are preserved and won't be overwritten on subsequent generations.

📊 Check the Schema Panel for Downstream

The column names in the Schema are critical. Downstream stages will reference these exact names in their transformations. Get them right!

📝 Review Node Configuration

Understanding the original node's configuration helps you validate the generated SQL. It shows inputs, parameters, and settings from your original flow.

Next Steps


Updated 4 days ago