You have your root environment for your new projects and the demo environment for your demo. Create a new environment and life is good. You use conda to create a new environment called demo that is built with whichever software packages and versions you need to run your demo. You break that code by reverting.Īlas, there is another option that solves all your problems. Not so fast! You’ve written code recently for another project that leverages new functions from the updated libraries.

If only you could revert your packages then all would be well, right? Because you’re whip smart, you enumerated package versions in your code so you know exactly what made it go many months ago. There is another option and it’s particularly enticing. You expected to devote a few minutes to knocking off the rust but now you’re contemplating reconfiguring old code. All in all this is not a great option either.Īt this point you’re starting to sweat. This likely requies some heavy lifting, though it could make your code a little more robust. Alternatively, you could write your own function to substitute in. Not the worst option but definitely requires a bit of effort without a guarantee of success. You could search for a substitute function in the new version. You get the infamous error message that some key function imported from Matplotlib was deprecated. So you dust off the old code and give it a run. Yay! You agree to run a one-hour session because, hey, the work is all done and you just need to brush up.

Lo and behold your old demo is the perfect substitute. Someone was supposed to give a demo at her conference but backed out last minute. No big deal, right?Ī friend of yours reaches out because she’s in a pinch.

You conda update -all and notice newer versions of NumPy and Matplotlib. Your code is beautiful, efficient, and the visualizations are mind boggling. For example, imagine you created the world’s most perfect Python demo that uses the NumPy and Matplotlib packages. Well it turns out these containers have a number of useful properties.

Simply put, you can think of a virtual environment as a container which houses software. What’s a virtual environment and why would you ever use one?

Creating conda environment