Anaconda Tutorial: Installing Specific Versions of Python Packages

In the previous tutorial, we learned how to search for available packages. Now, we’ll delve deeper into specifying the exact version of a package to install. When managing packages with Anaconda, you may occasionally need a particular package version—either to meet project-specific requirements or ensure compatibility across dependencies. This article demonstrates, through multiple practical examples, how to use the conda command to install packages at specific versions.

Installing a Package at a Specific Version

In Anaconda, you can install a package at a precise version using the conda install command. The basic syntax is:

conda install package_name=version_number

Here, package_name is the name of the package you wish to install, and version_number is the exact version you want.

Example 1: Installing a Specific Version of NumPy

Suppose your code relies on NumPy version 1.18.5 to run correctly. You can install that version with the following command:

conda install numpy=1.18.5

After executing the command, conda automatically resolves dependencies and prompts you to confirm the installation. Once confirmed, conda downloads and installs the specified NumPy version.

Important Notes

During installation, if another version of the same package is already installed, conda handles version conflicts—potentially removing the existing version to accommodate the requested one. To check which version is currently installed, use:

conda list numpy

Installing Multiple Packages at Specific Versions

You can also install several packages—each at a designated version—in a single command. For instance, suppose you need Pandas version 1.1.5 and Matplotlib version 3.3.2. Use:

conda install pandas=1.1.5 matplotlib=3.3.2

Example 2: Installing Multiple Packages at Specific Versions

Imagine setting up a data analysis environment requiring specific versions of Pandas and Matplotlib:

conda install pandas=1.1.5 matplotlib=3.3.2

With this command, conda resolves inter-package dependencies and ensures all selected versions are compatible and co-installable.

Installing Packages Within a Version Range

Anaconda also supports installing packages within a defined version range. For example, you might want any Scikit-learn version between 0.24 (inclusive) and 0.24.2 (exclusive). Specify it like this:

conda install scikit-learn>=0.24,<0.24.2

Example 3: Installing a Package Matching a Version Range

This command instructs conda to select and install the highest-compatible version of Scikit-learn satisfying the constraint—and automatically resolve its dependencies:

conda install scikit-learn>=0.24,<0.24.2

Listing Available Versions

Before installing a specific version, it’s helpful to know what versions are available. Use the following command to list all available versions of a given package:

conda search package_name

For example, to see all available versions of SciPy, run:

conda search scipy

This command displays a comprehensive list of versions—including build information and channel sources—to help you make an informed choice.

Summary

In this tutorial, we covered how to use conda to install Python packages at specific versions. Through concrete examples, we demonstrated installing individual packages, multiple packages simultaneously, and packages matching flexible version constraints. In upcoming tutorials, we’ll explore installing packages from different channels—highlighting Anaconda’s flexibility and robustness as a package management system.

We hope these examples deepen your understanding and confidence in using Anaconda for effective package management. If you have questions or feedback, feel free to leave a comment below!