How to install multiple python packages at once using pip

Customized installations

Often times, there can be a dependency on the versions of packages being used. In this case, you can use a requirements.txt file like this:

# "Orders up" for the barista at the pip cafe!
numpy==1.19.3
pandas>=1.1.5,<1.2.0
requests>=2.25.0

Execute version-controlled installation:

pip install -r requirements.txt


The version of numpy installed will be exact, for pandas it will be the latest version which is 1.2.0 or lower, and for requests, it will be 2.25.0 or higher.

Nifty production tips

For production environments, it pays to consider the following when preparing a requirements.txt file:

• Version pinning to avoid unexpected updates (use ==).
• Specify compatible versions (use <=, >=, !=, ~=).
• Test with the specified versions before deploying to the production environment.

This creates a predictable, stable and reliable requirements.txt which ensures a smooth roll-out and stability post-deployment.

Harness the power of a requirements file

Wouldn't it be great if you could generate a requirements.txt file from an existing environment? As easy as:

# Go-go gadget "freeze"! 
pip freeze > requirements.txt


This command neatly stacks every installed package, along with its buttery smooth version, into a text file. You'll find it incredibly useful when:

• Migrating environments: Like importing your grandma's secret pie recipe into your new kitchen.
• Version locking: Keeping things as they were. No unexpected "suprise features"!
• Project documentation: Giving your teammates a map to the bounty of quick setup.

Gracefully handling installation issues

Should you encounter hiccups, here are some typical remedies:

• Connectivity: Check your internet. Packages don't teleport (yet)!
• Spelling: Don't get tripped up missing 'i's or mixing 'e's. Package names are often crafty.
• Virtual environment: Best practise. Keeps everything neat and tidy. No permission issues, no conflicts.

Creating a virtual environment is as easy as pie:

python3 -m venv myenv
source myenv/bin/activate


Then, just run pip commands in this isolated space. Neat, right?