Cloudera Docs

Tutorial: Building a new flow from scratch

If you are new to flow design and have never used NiFi before, this tutorial is for you. Learn how to build a draft adding and configuring components, connecting them, creating Controller Services, and testing your flow while creating it.

This tutorial walks you through the creation of a simple flow design that retrieves the latest changes from Wikipedia through invoking the Wikipedia API. The flow converts JSON events to Avro, then filters and routes the events to two different processors which merge events together, and finally a file is written to local disk.

You will learn about the following actions:

  • Creating a draft
  • Creating a Controller Service
  • Adding processors to your draft
  • Configuring processors
  • Adding a user-defined property to a processor configuration
  • Connecting processors to create relationships between them
  • Running a Test Session
  • Publishing a draft to the Catalog as a flow definition

The flow you are about to build can be deployed without any external dependencies and does not require any parameter values during deployment. Still, you must meet the following prerequisites before you can start building your first draft:

  • You must have an enabled and healthy Cloudera Data Flow environment.

  • You must be assigned the DFDeveloper role granting you access to the Flow Designer.

  • You must be assigned the DFCatalogAdmin or DFCatalogViewer role granting you access to the Catalog. You must have this authorization to publish your draft as a flow definition to the Catalog.

  • You must be assigned the DFFlowAdmin role for the environment where you want to deploy the flow definition.