Python Patterns: max Instead of if
I often have to loop over a set of objects to find the one with the greatest score. You can use an if statement and a placeholder, but there are more elegant ways!
When writing Python, I often have to look through a set of objects, determine a score for each one of them, and save both the best score and object associated with it. For example, looking for the highest scoring word that I can make in Scrabble with the letters I currently have.
One way to do this is to loop over all the objects and use a placeholder to remember the best one seen so far, like this:
You have probably written this logic before in some of your own code. The code is not that complicated, but we can still improve its readability with a quick tweak.
if score > best_score remind you of? The way we might implement the
max() function! Using
max() helps us simplify the code nicely:
Storing all the data together in a single tuple means that assignment and comparison are now handled all at once. This makes it less likely that we will mix up one of the assignments, and makes it clearer what we’re doing.
There is one potential pitfall here:
max() picks the tuple with the largest first element (the score in our case), which is what we want. But, if the first elements are the same in both tuples,
max() continues through the remaining elements until the tie is broken. So if two words have the same score,
max() will then compare the words next, which it does lexically.
max() only compare the first element, we can use the
key parameter. The
key parameter takes a function that is called on each object and returns another object to use in the comparison. We can use it to select just the first entry like so:
In the above examples we wanted to save both the score and the word, but what if we only cared about the word that generated the highest score, not the score itself? Then there is an even simpler way!
max() uses the standard comparison operator, but we can change that to use our
score_word() using the same
key argument from above. Then we have:
Which gives us a very compact (and relatively fool proof) pattern, with all the looping and placeholders pushed into the implementation of
Bio: Alexander Gude is currently a data scientist at Intuit using machine learning for fraud prevention. He holds a BA in physics from University of California, Berkeley, and a PhD in Elementary Particle Physics from University of Minnesota-Twin Cities.
Original. Reposted with permission.
- 5 “Clean Code” Tips That Will Dramatically Improve Your Productivity
- How Different are Conventional Programming and Machine Learning?
- Get a 2–6x Speed-up on Your Data Pre-processing with Python