CSVs were and still are a very popular format for importing and

exporting data to and from databases and spreadsheets.

Much of the web though, runs on another common and

popular format, JSON, or JavaScript Object Notation.

No, I'm not trying to get you to write JavaScript,

Python's JSON module is very handy for encoding and decoding data from JSON.

You'll use this when you're receiving information in your API or

sending information to an API, or maybe you just want to use it to read and

write files on your server.

JSON is pretty common for

configuration files too, regardless of how you're going to use it,

it's a great format to know about, and a great library to know how to use.

Python's JSON library is pretty solid and I think, pretty easy to use.

Let's check out this JSON file about a piece of art

that I downloaded from the Minneapolis Institute of Art.

You can see that a lot of stuff in here, I can read it like it's a dictionary and

go, it's dated 1899, cool, but I don't wanna do that.

I wanna be able to do this in Python, right?

So I'm gonna make a new file and I'm gonna call this Art.pi,

and inside of here, first things first.

We need to import the JSON module, so

the JSON module has all of our handy dandy stuff in it.

So I'm going to open it up, what is this?

148.json as art file.

So I'm gonna open that up and I'm gonna say art = jason.load(artfile),

and then I'm gonna print(art['description']).

All right, cool?

So let's go run this.

And I get that it's a boat with two masts in front of a bridge with a lift-gate,

two row boats in water at left, buildings at left, trees at right,

that's the description I got from the JSON file.

Now that might not look like much but it's really handy if you're dealing with tons

of input from an outside source, or lots and lots and lots of files.

And you don't have to try and turn plain-text files in Python Objects or

anything like that through RegX or some sort of, I don't even know,

some sort of arcane incantation.

The load method, which we used right here, takes a file-like object and

tries to turn that into a Python object of some kind.

In this case, because what we have here is what we would call a dictionary in Python,

it just creates a dictionary, it creates an actual Python dictionary.

If we had a list, an array for JavaScript, then we would get a list

in Python, another way that we can do this is with the load s method.

You'll probably hear people call it loads instead of load s or

load string, I call it loads all the time.

Honestly, that's fine, that's what the word is, but the way that it's spelled,

what it is, is it's the way that we can take a string with embedded JSON in it and

turn that into a Python object.

Let's try that out, let's try it out down here in the shell.

We'll do Python, import json and

let's do nums = json.loads and

then a string of 1,2, and 3.

All right, and so now if I do nums[2] I get the number 3,

so we loaded the string in the JSON.

In that string there was an array of integers which the JSON module turned

into a list of ints and I think that's pretty handy.

The opposite of load and loads are the dump and

dumps methods, so let's try them out.

So we can do json.dumps([5, 4, 3, 2, 1]).

No quote marks, and now, I get a string which has the array of numbers in it,

so it's kinda like I turned the list into a string,

I mean, it almost seems like I get to str and then the list, right?

And I could on something this simple, but on something that's a little bit

more complicated, you obviously can't do that, let's try this.

I'm gonna make myself into a dictionary, so first_name is Kenenth,

last_name is Love, and my topic here at Treehouse Is Python.

All right, so I can't just do str(me),

I mean, I can but I don't get exactly the right thing, right?

So let's do json.dumps(me) and

while it looks the same, there is a difference because in JSON

you have to have double quotes instead of single quotes to make it valid.

So this is valid JSON, what I did before is not, and if I was to have

numbers like that in there, we'd want to change how those things worked.

We could also easily send this string out across the Internet with no problems, so

that's great,

that's super handy, we could do something similarly using the Dump method.

But Dump goes directly into a file-like object instead of into a string,

so we have me, let's make Craig.

So again, do first_name as Craig,

last_name as Dennis, and topic as Java all right?

So we have me, and we have Craig, and let's do with open

teachers.json, and a for append, although it doesn't really matter,

cuz we're writing into this for the first time, as teacherfile.

Then we wanna say json.dump, and

we're gonna dump an array of me and

craig into teacherfile, all right?

Let that run and now, let's shrink that down, refresh this,

if we checked out teachers.json that the file that we just created.

There's me, and I misspelled my own name, and

there's Craig, and it made a list of objects.

So that's pretty cool, that's super awesome, the JSON module is amazingly

handy for both encoding and decoding data to and from files and strings.

I find myself using it a lot with various APIs on the web, and

I bet that you will find it the same.