Stricter Zipping

00:00 In the previous lesson, you learned about the three new functions in the statistics module. In this lesson, I’ll show you the zip() function and its new strict parameter.

00:10 The built-in zip() function is used to combine two or more sets of data into a set of tuples. The easiest way to understand this is to use an example.

00:19 Consider this table that shows data for five different sets of Lego. Each set has a name, an ID number, and some number of pieces. I’ll use this data in the REPL to show you zip().

00:33 Let’s say you’ve got three lists, one for each column in the Lego table. What zip() lets you do is create a series of tuples, each tuple containing the name, number, and size of a Lego set.

00:46 You run zip() by passing in the sequences you want to operate on. And you get back a generator. That’s great from a memory management perspective but not particularly instructive. So let me do that again, converting the generator into a list.

01:11 That’s better. The result of the zip is a list of tuples, each tuple being about a Lego set—from the famous French museum to the New York City skyline. Now, here’s the footgun.

01:24 zip() just assumes your data is all good, but what if it isn’t? Let me remove the last number from the ID set.

01:38 Now, set_numbers has only four items in it. The name and pieces lists still have five members, but the set ID sequence now only has four.

01:56 Running zip() on this corrupted data is problematic. It just does it. This can be a hard bug to find. You end up with a zipped sequence with just one less item in it. This is even worse if the missing piece of data is in the middle.

02:09 The wrong things will get associated with each other. Python 3.10 to the rescue.

02:22 zip() now supports an optional argument called strict. When set to True, zip() will no longer work if there is a length mismatch. It throws a ValueError instead.

02:34 There is an alternative to this that’s been kicking around for a while in the itertools module. It’s called zip_longest(). It works like zip() but inserts None into the data when there’s a length mismatch. Let’s run it on the Lego data.

02:56 This time, New York City is included, it just ends up with None as its set ID. zip_longest() also takes an optional argument called fillvalue.

03:06 This allows you to fill in the empty data with something other than None, that you specify.

03:18 Here, I’ve set fillvalue to 0, and you can see the difference at the end of the sequence in the New York City tuple. Appropriately enough, the lesson on zip() was rather zippy. Next up, a miscellaneous, catchall, collage, grab bag, hodgepodge, potpourri smorgasbord of the smaller stuff that’s left in 3.10. Why, I do own a thesaurus! Why do you ask?