KNIME Documentation

The Pixi Backend ​

The modern Python integration in KNIME is powered by Pixi, a package management tool built on the Conda ecosystem.

Key Benefits of the Pixi Backend ​

  • Cross-platform by design: Environments are resolved for all target operating systems (Windows, Linux, macOS) simultaneously.
  • Fully reproducible: The exact state of the environment is saved as a pixi.lock file, guaranteeing identical package versions everywhere.
  • Significantly faster: Both environment resolution and node configuration are much faster. The node dialog opens instantly, eliminating the 2-minute wait times associated with legacy Conda nodes.
  • Zero setup required: Pixi comes pre-installed and encapsulated within KNIME. No manual Conda installation or user configuration is needed.
  • In-workflow definition: You can define and edit the environment directly inside the KNIME workflow. In contrast, the legacy Conda Environment Propagation (CEP) node could only read pre-created local environments.
  • Browser-ready modern UI: The node's modern interface is designed to work seamlessly in the browser for future cloud-based editing features.

When to use which system ​

Choosing between the modern Pixi-based nodes and legacy Conda-based setups depends on your deployment needs and environment access.

Use the Python Environment Provider (Modern/Pixi) if: ​

  • You are deploying to the Hub: This is the recommended way to guarantee that an environment built on Windows will resolve and install correctly on a Linux-based Hub executor.
  • You want "Zero Setup": Use this to avoid the manual installation and configuration of Conda or Miniforge in the KNIME Preferences.
  • You need specific versions per workflow: Since the environment is defined at the node level, you can have different workflows using conflicting library versions without interference.
  • You are preparing for cloud editing: While browser-based editing is a future feature, using these nodes now ensures your workflows will be compatible once released.

Use Conda / Manual Configuration (Legacy) if: ​

  • You have a pre-existing, complex local environment: If you have a large, curated Conda environment on your local machine that is difficult to replicate via a simple specification.
  • You have restricted internet access: If your security environment blocks Pixi from reaching external repositories and you haven't configured a local mirror or a Conda fallback.
  • You need a specific system-level Python: If your code must interact with a specific Python binary installed at the OS level (e.g., for hardware-specific drivers).

Key Concepts ​

Port Objects vs. Flow Variables ​

The legacy system passed environment information through Flow Variables (red connections). The modern integration uses Port Objects (square connections).

Environment Locking ​

The "Lock Package Versions" feature in the PEP node performs a compatibility check across all target platforms simultaneously. It generates a deterministic lock-file (pixi.lock) that guarantees every executor, whether on a laptop or a Hub instance, installs the exact same version of every library.