ZipponDB/README.md

ZipponDB

Note: Make a stupide mascotte

Written in Zig

Zig is fast, blablabla

How it's work

Meme "That's the neat part..."

Zippon is a strutural relational potentially in memory database written entirely in Zig from stractch.

You build a binary according to your schema, you can just run it to acces a CLI and it will create and manage a folder 'zipponDB_DATA'. Then you do what you want with it, including:

• Run it with your app as a seperated process and folder
• Create a Docker and open some port
• Create a Docker with a small API like flask
• Other stuffs, Im sure some will find something nice

Integration

Python

import zippondb as zdb

client = zdb.newClient('path/to/binary')
print(client.exe('describe'))

# Return named tuple
users = client.exe('GRAB User {}')
for user in users:
    print(user.name)

client.exe('save')

Benchmark

I did a database with random data. The schema is like that:

User {
    name: str,
    email: str,
    friends: []User.friends,
    posts: []Post.from,
    liked_post: []Post.like_by,
    comments: []Comment.from,
    liked_com: []Comment.like_by,
}

Post {
    title: str,
    image: str,
    at: date,
    from: User.posts,
    like_by: []User.liked_post,
    comments: []Comment.of,
}

Comment {
    content: str,
    at: date,
    from: User.comments,
    like_by: User.liked_com,
    of: Post.comments,
}

As you can see, link need to be defined in both struct. [] mean an array of value. For example posts: []Post.from, and from: User.posts, mean that a User can have multiple posts (an array of Post) and a post just one author. Both linked by the value posts and from.

Create a schema

Zippon use struct as way of saving data. A struct is a way of storing multiple data of different type. Very similar to a row in a table, columns being datatype and a row a single struct.

The schema is directly INSIDE the binary, so each binary are per schema ! This is for effenciency, idk to be honest, I guess ? lol

Migration

For now you can't migrate the data of one binary to another, so you will need to different binary.

Zippon language

Ok so I went crazy on that, on have it how language. It is stupide and I love it. I wanted to do like EdgeDB but no, too simple. Anyway, I tried to do something different, to do something different, idk, you're the jduge of it.

GRAB User { name = 'Adrien' }
Get all user named Adrien

GRAB User [1; email] { }
Get one email

GRAB User {} | ASCENDING name |
Get all users ordered by name

GRAB User [name] { age > 10 AND name != 'Adrien' } | DECENDING age |
Get just the name of all users that are 10 years old or more and not named Adrien ordered by age

GRAB User { bestfriend = { name = 'Adrien' } }
GRAB User { bestfriend = User{ name = 'Adrien' } } // Same
Get all user that have a best friend named Adrien

GRAB User [10] { IN User [1] { age > 10 } | ASC name |.friends }
Get 10 users that are friend with the first user older than 10 years old in ascending name order

GRAB Message [100; comments [ date ] ] { .writter = { name = 'Adrien' }.bestfriend }
Get the date of 100 comments from the best friend of the writter named Adrien

GRAB User { IN Message { date > '12-01-2014' }.writter }
Get all users that sended a message after the 12 january 2014

GRAB User { !IN Comment { }.writter }
Get all user that didn't wrote a comment

GRAB User { IN User { name = 'Adrien' }.friends }
Get all user that are friends with an Adrien

UPDATE User [1] { name = 'Adrien' } => ( email = 'new@email.com' )

REMOVE User { id = '000-000' }

ADD User ( name = 'Adrien', email = 'email', age = 40 }

• {} Are filters
• [] Are how much; what data
• () Are new or updated data (Not already savec)
• || Are additional options
• Data are in struct format and can have link
• By default all value other than a link are return per query, to prevent recurcive return (User.friends in User.friends)

How it's really work

NOTE: Do this in a separe file

Tokenizer

The tokenizer of the language is

ZipponDB