Create and Use a Conda Environment as the Project Interpreter in JetBrains Dataspell

Desired outcome

You want an isolated Conda environment containing the exact versions of libraries (e.g., numpy, pandas, scikit-learn) that your Dataspell project will use. Once configured, Dataspell will use that environment for code completion, debugging, and the Python console, giving you a reproducible workflow.

Prerequisites

• Dataspell version 2023.2 or later installed.
• Conda (Anaconda or Miniconda) installed and its conda executable accessible from the system PATH.
• A Dataspell project opened or created.

Procedure

1. Open the Settings/Preferences dialog: Ctrl+Alt+S (Windows/Linux) or ⌘+, (macOS).
2. Navigate to Project: <your‑project‑name> → Python Interpreter.
3. Click the gear icon ► Add….
4. In the left pane select Conda Environment.
5. Choose the New environment tab.
6. Fill in the fields:
   • Name: e.g., ml-env
   • Base interpreter: leave as the detected Conda Python (or browse to conda if needed).
   • Location: accept the default or specify a local SSD path for better performance.
   • Python version: select 3.11 (or another version required by your project).
7. Optionally check Inherit global site‑packages if you need access to packages installed in the base Conda environment; otherwise leave it unchecked for full isolation.
8. Optionally check Make available to all projects if you plan to reuse this environment elsewhere.
9. Click OK. Dataspell will create the environment and then display it in the interpreter list.
10. Select the newly created environment from the interpreter dropdown and click Apply.
11. To install additional packages via the UI, open View → Tool Windows → Python Packages, ensure the environment is selected, and search/install numpy, pandas, scikit-learn.
12. (Alternative) You can also create the environment from a terminal and then point Dataspell to it:

    conda create -n ml-env python=3.11 numpy pandas scikit-learn
    conda activate ml-env
    # verify
    python -c "import sys, numpy, pandas, sklearn; print('OK')"
    

    After creation, repeat steps 2‑5 but choose Existing environment and browse to <conda‑envs‑path>/ml-env.

Expected checks

• The interpreter dropdown in Settings → Project → Python Interpreter shows the new environment (e.g., ml-env).
• Open the Python console from the bottom toolbox and run:

  import sys; print(sys.executable)
  

  The output should point to the python.exe (or python binary) inside the environment’s directory.
• Run:

  import numpy, pandas, sklearn; print(numpy.__version__, pandas.__version__, sklearn.__version__)
  

  to confirm the expected packages are present.
• In the Python Packages tool window, the list should include numpy, pandas, and scikit-learn with the versions you installed.

Recovery options and limitations

• If the environment does not appear in the interpreter list, verify that Conda is on the PATH by opening a system terminal and running conda info. The command should return version and environment information without error.
• Invalidate Dataspell caches via File → Invalidate Caches / Restart if the UI still does not reflect the new environment.
• Creating the environment on a network‑mounted drive can slow down package indexing and code completion; prefer a local SSD location.
• Avoid mixing conda install and pip install within the same environment unless you use pip from the activated Conda context (conda activate ml-env && pip install …). Mixing managers can lead to inconsistent package states.
• Changing the interpreter requires restarting any running Python console or debug sessions to use the new environment.