Skip to main content
After successfully executing the workflow, it’s now time to get the outputs and results:
  1. ✅ Import nodes and builder
  2. ✅ Create the workflow builder
  3. ✅ Define and add nodes
  4. ✅ Connect the nodes
  5. ✅ Build and visualize the pipeline
  6. ✅ Execute the pipeline
  7. ➡️ Get outputs and results
For both ✍️ Example 1 and 📄 Example 2 from Create Workflow section, here is how to execute workflows that have been built previously:
  1. ➡️ Get outputs and results
# Execute pipeline and get resuls
results = pipeline.execute()

# Show results
display(results)  # or simply `results` at the end of a notebook cell
output: 
{
		'data':
				FlowData(...),
		'error': None,
		'node_results': {
				0: NodeResult(node_id=0, data=FlowData(...),
				1: NodeResult(node_id=1, data=FlowData(...),
				2: NodeResult(node_id=2, data=FlowData(...),
				...
		}            
}

​
Explore results

  • results.data : contains the state of FlowData object after the execution of the last node
  • results.node_results : dictionary containing the states of FlowData object after the execution of each node of the workflow. Using this field, we can traceback the intermediate steps of the workflow
  • results.error : if an error occurs during workflow execution, the error message will be saved in this field

​
Remarks

  • Workflow always goes to the end, even if an error occurs during the execution.
  • If no error occurred during execution, results.error is None
  • If an error occurred during execution, results.error contains the error message
⌘I