The new case_when method in Pandas 2.2.0

Pandas 2.2.0 includes a new method, case_when. What is it, and how can you use it? I give an overview, with a number of examples.

Here's a simple example of passing a callable, albeit not the most interesting one:

(lambda s_: s_.ge(500), 'high')])

The conditions here are lambdas (aka anonymous functions) that are invoked by case_when, rather than boolean series.

[imothar]

The info about default value was very helpful for me. Thank you👍

[christophertyler]

Reminds me of the case statement in SQL. I can see multiple situations that this will simplify my Pandas code, thanks for sharing!

[datossinmisterios]

great video! i still don't understand how you don't have at least 500k followers!

[robertrussell]

This is a good addition to Pandas functionality.

[ssss]

Thanks again for this video, it's always helpful. As a course suggestion: how to optimise the computation of data using the previous values in the series/df, especially when we need the result of the computation of the previous rows (i.e. multiprocessing not possible as you have to iterate the whole series in order).

[stefaandumez]

Why do you import Series from pandas. Isn't that part of the whole package, already imported? Great content btw!

[ofirmazor]

This is nice but not so different than using np.where().
I belive np.where is much faster when dealing with big serieses.