jdbl

jdbl

A json based database library, RAM running and including basic read-write lock.

TODO:

• add documentation

What is jdb?

jdb means json based database, but with some fixed structure.

database & table

jdbl made its own file format, .jdb.

A .jdb file's data is a database(db) itself. Its each key is a table(tb) name, and the value is a table data.

Seems like this:

// db1.jdb
{
    "tb1": {...},
    "tb2": {...}
}

table data

As said above, the value of each key of jdb is a table data. A table data is a json format data too, with two fixed keys: columns and rows.

The value of columns is a list of columns' infos, and the value of rows is a list of rows' datas.

Seems like this:

// db1.jdb
{
    "tb1": {
        "columns": [...],
        "rows": [...],
    },
    "tb2": {
        "columns": [...],
        "rows": [...],
    }
}

For understanding better, let's see rows structure first.

rows

rows is a list of row data, it is a list of dict. And each row data is a dict, with attribute name as key, and attribute value as value. You can think a row data is descrbibing like: A person with name "Alice", age 18.

So rows datas seems like this:

...
    "rows": [
        {"id": "32132654", "name": "Alice", "age": 18, ...},
        {"id": "65653251", "name": "Jerry", "age": 19, ...},
        ...
    ],
...

Each attribute name would have a corresponding column info in columns. As you can see below.

columns

columns is a list of column infos, it is a list of dict too. Each column info is a dict, with column's info name as key, and column's info value as value.

There are two fixed info names are fixed.

• name: the name of the column.
• rname: the readable name of the column.

And any other info names are custom info names, you can define them by yourself. The jdbl library would make sure that every row data attribute name is in the columns list.

Seems like this:

...
    "columns": [
        {"name": "id", "rname": "ID", ...},
        {"name": "name", "rname": "Person's name", ...},
        {"name": "age", "rname": "Person's age", ...},
    ],
...

How to use

This jdbl library helps handle .jdb file, and it provides some functions to operate the database.

install & import

You can install it by pip:

pip install jdbl

And import it:

from jdbl import jdb_handler

Lock & Unlock

In order to avoid data corruption for multi-threaded, multi-processed environment, jdbl provides read-write lock directly. It means when you used(read or write) one jdb file, the file would be locked up. To make the api more readable, the function is named use and un_use.

use & un_use

use function is used to lock the jdb file, and un_use function is used to unuse the jdb file.

# lock the jdb file
jdb_handler.use("db1")

# unlock the jdb file
jdb_handler.un_use("db1")

When you use a jdb, you should make sure that you unlock it after you finish using it. Otherwise next time you use the same jdb, it would be block for five seconds and print a warning.

Query Datas

As it is a json format with fixed structure, you can use json library to get datas directly. But jdbl provides some functions to make it more readable.

get_tb_data

get_tb_data function is used to get the table data of a table.

tb_data = jdb_handler.get_tb_data("db1", "tb1")

query_rows

query_row function is used to query the rows that with some conditions. For now, it searches rows by col_name and col_value provided.

rows = jdb_handler.query_rows("db1", "tb1", "name", "Alice")

This would return a list of row data dict that with name attribute value is "Alice".

Write Datas

As it is a json format with fixed structure, you can use json library to write datas directly. But jdbl provides some functions to make it more readable.

add_rows

add_rows function is used to add rows to a table. And it has upsert mode. Which would just update the row if it exists, or add the row if it doesn't exist.

row_datas = [
    {"id": "32132654", "name": "Alice", "age": 18, ...},
    {"id": "65653251", "name": "Jerry", "age": 19, ...},
    ...
]
jdb_handler.add_rows("db1", "tb1", row_datas, "name", upsert_mode=True)

This would add the rows to tb1 table in db1 database. If the row with name attribute value is "Alice" exists, it would be updated. Otherwise it would be added.

One-time functions

As said above, normally you need to use a jdb and do something with it, then you need to un_use it. It is for fast RAM running when you need to do many operations to one jdb. But for some one time operation like "reading the table and do nothing else", the use-do-un_use steps are a little bit annoying. So jdbl provides some one-time functions to make it more convenient.

Here're some one-time functions like:

• get_tb_data_once
• query_rows_once
• add_rows_once

They are the same as the normal functions, but you don't need to use and un_use the jdb yourself, the function would do it for you.