Python 3.15: Exploring the New Features

Yeah, I want to start out with something that solves a very specific problem.

This will likely be the part of the release that's actually most useful to me, but I suspect I might be a minority, but they've got at least one fan.

I'll get to the feature in a minute.

Hopefully, I'll be able to make it clearer with the use case.

So I'm going to start out with that.

Say you've got a web application where you're shipping some JSON back and forth that describes an object.

When the user changes something, a new bundle of JSON gets sent to the server.

There are several ways of doing this, but REST is probably the most popular.

REST uses HTTP methods, and one of those methods is PUT.

When you call PUT, you send along all of the data that describes the object in the browser to the server.

So if you were on, say, a change address form or shipping form, whatever, you'd send all of the parts of the address, the address lines, the city, the zip code, all that kind of good stuff in one JSON bundle over, including not only the part that changed, but everything.

So as you might guess, this is a bit expensive.

For small amounts of data, it doesn't really matter.

But if you're dealing with a large set of objects, it's overkill to send the entire object state if only one field has changed.

So there is another REST method that is more efficient, and that's HTTP patch.

In that case, you only send the changed field up to the server.

So on the server side, you typically run the data through a validator because you never trust the browser because the browser and people are evil.

So you might use a tool like Pydantic and a Pydantic object has all of the fields that the client side does.

But when you're only editing one field using a patch, it sets the ones that didn't get touched to none.

So your code then loops through the attributes on the object and changes the server side data based on what attributes are set and ignoring anything that's set to none.

Where this gets messy is when you delete a field on the client side.

So if it's a string, you can set it to the empty string.

But what if it's an integer? You could send the default value, but sometimes you want to distinguish between default and actually empty.

The way to do this in JSON is to set the field to null, which is JavaScript's equivalent of none.

And this is where the problem comes from.