mql2

MQL - Node.js Functional SQL Query Builder & ORM

EN | KR

Features

• Tagged template literal
• No models.
• Only need functions and javascript data types.
• Promise.
• No cost for converting to JSON.
• More freedom in using SQL syntax.
• Preventing SQL-injection attacks.
• Easy to use the latest operators provided in databases.
• Simple transaction API.
• No models for Associations.
• Designed to work well with PostgreSQL, MySQL.

Overview

• Installation
• Connect (pool)
  • PostgreSQL
  • MySQL
• Closing all the connections in a pool
• Simple query
• Subquery, Join
• Ready to be used
• Helper Function
  • EQ
  • IN
  • NOT_IN
  • VALUES
  • SET
  • COLUMN, CL
  • TABLE, TB
• Associations
  • Common use
  • Polymorphic
  • Transaction
  • Many to many
  • ROW_NUMBER + PARTITION
  • Hook
  • ASSOCIATE_MODULE
• Option
• DEBUG

Installation

npm i mql2

Connect (pool)

PostgreSQL

const { PostgreSQL } = require('mql2');
const { CONNECT } = PostgreSQL;
const POOL = CONNECT({
  host: 'localhost',
  user: 'username',
  password: '1234',
  database: 'dbname'
});

PostgreSQL Connection option

MQL is built on node-postgres. The parameter of CONNECT function is the same as node-postgres’. You can read the detail of connection pool or connecting to DB on node-postgres’ site.

MySQL

const { MySQL } = require('mql2');
const { CONNECT } = MySQL;
const POOL = CONNECT({
  host: 'localhost',
  user: 'username',
  password: '1234',
  database: 'dbname'
});

MySQL Connection option

MQL is built on node-postgres. The parameter of CONNECT function is the same as the MySQL’. You can read the detail of connection pool or connecting to DB on MySQL's site.

Closing all the connections in a pool

PostgreSQL, MySQL

POOL.END(); // Promise

Simple query

const { QUERY } = POOL;
const id = 10;
const posts = await QUERY `SELECT * FROM posts WHERE id = ${id}`;
// [{ id: 10, ... }]

Subquery, Join

const type = 'TYPE1';
const limit = 10;

QUERY `
  SELECT * FROM table1 WHERE table2_id IN (
    SELECT id FROM table2 WHERE type = ${type} ORDER BY id DESC LIMIT ${limit}
  )
`;

const status = 'STATUS1';

QUERY `
  SELECT *
    FROM table1 AS t1, table2 AS t2
    WHERE t1.id = t2.table1_id AND t1.status = ${status}
    ORDER BY id DESC
    LIMIT 10
`;

QUERY achieved from CONNECT uses a connection pool.

Ready to be used

const POOL = CONNECT();
const {
  VALUES, IN, NOT_IN, EQ, SET, COLUMN, CL, TABLE, TB, SQL, MQL_DEBUG,
  QUERY,
  ASSOCIATE,
  LJOIN,
  TRANSACTION,
  END
} = POOL;

Helper-Function

EQ

const users = await QUERY `SELECT * FROM users WHERE ${EQ({
  email: '[e-mail]',
  password: '1234'
})}`;
// [{ id: 15, email: '[e-mail]', ... }]

IN

const users = await QUERY `SELECT * FROM users WHERE ${IN('id', [15, 19, 20, 40])}`;
// [{ id: 15, ...}, { id: 19, ...} ...]

NOT_IN

const users = await QUERY `SELECT * FROM users WHERE ${NOT_IN('id', [2, 4])} ORDER BY ID LIMIT 3`;
// [{ id: 1, ...}, { id: 3, ...}, { id: 5, ...}]

VALUES

const post = { user_id: 10, body: 'hoho' };
await QUERY `
  INSERT INTO posts ${VALUES(post)}
`;
// INSERT INTO posts ("user_id", "body") VALUES (10, 'hohoho')

await QUERY `
  INSERT INTO coords ${VALUES([
    {x: 20},
    {y: 30},
    {x: 10, y: 20}
  ])}`;
// INSERT INTO coords ("x", "y") VALUES (20, DEFAULT), (DEFAULT, 30), (10, 20)

SET

await QUERY `
  UPDATE posts ${SET({ body: 'yo!', updated_at: new Date() })} WHERE id = ${post.id}
`;
// UPDATE posts SET "body" = 'yo!', "updated_at" = '2018-08-28T23:18:13.263Z' WHERE id = 10

COLUMN, CL

COLUMN == CL; // true

await QUERY `
  SELECT
    ${COLUMN('id', 'bb as cc', 't2.name', 't2.name as name2', { a: 'c' }, { 't3.a': 'd' })}
      ...
`;
// SELECT
//   "id", "bb" AS "cc", "t2"."name", "t2"."name" AS "name2", "a" AS "c", "t3"."a" AS "d"
//     ...

TABLE, TB

TABLE == TB; // true

await QUERY `
  SELECT
    ...
    FROM ${TABLE('t1')}, ${TABLE('tt as t2')}
`;
// SELECT
//   ...
//   FROM "t1", "tt" AS "t2"

Associations

Common use

ASSOCIATE uses Connection pool.

/*
* users
*  - id
*  - name
*
* posts
*  - id
*  - user_id
*  - body

* comments
*  - id
*  - user_id
*  - post_id
*  - body
* */

const { ASSOCIATE } = POOL;

const posts = await ASSOCIATE `
  posts
    - user
    < comments
      - user
`;

posts[0].body;
posts[0]._.user.name
posts[0]._.comments[0].body
posts[0]._.comments[0]._.user.name

- of - user refers to "Belongs to", < of < user refers to "Has many".

Polymorphic

/*
* photos
*  - attached_type
*  - attached_id
* */

await ASSOCIATE `
  posts
    - user
      p - photo
    p < photos
    < comments
      p < photos
`;
// SELECT * FROM photos WHERE attached_id IN (${map($ => $.id, posts)}) AND attached_type = 'posts';
// SELECT * FROM photos WHERE attached_id IN (${map($ => $.id, users)}) AND attached_type = 'users';
// SELECT * FROM photos WHERE attached_id IN (${map($ => $.id, comments)}) AND attached_type = 'comments';

p - refers to Polymorphic + Has one, p < refers to Polymorphic + Has many.

Many to many

/*
* books
*  - id
*  - title
*
* authors
*  - id
*  - name
*
* books_authors
*  - author_id
*  - book_id
* */

const books = await ASSOCIATE `
  books
    x authors
`;

books[0]._.authors[0].name;

const authors = await ASSOCIATE `
  authors
    x books ${{ xtable: 'books_authors' }}
`;

authors[0]._.books[0].title;

Option

/*
* If the tables are formed like the example below, the ASSOCIATE automatically creates the necessary table and column names for queries. the necessary names for the tables and columns for queries
* users
*  - id
* posts
*  - id
*  - user_id
* comments
*  - id
*  - post_id
*  - user_id
* likes
*  - attached_type
*  - attached_id
*  - user_id
* posts_tags
*  - post_id
*  - tag_id
* tags
*  - id
* */

ASSOCIATE `
  posts
    - user
    < comments
     - user
     p < likes
      - user
    p < likes
      - user
    x tags
`;

/*
* You can select columns or add conditions.
* Even though you don’t select a foreign key or a primary key in the option like the below, they are included in ASSOCIATE.
* */

ASSOCIATE `
  posts ${SQL `WHERE is_hidden = false ORDER BY id DESC LIMIT ${10}`}
    - user
    < comments ${{
      column: COLUMN('body', 'updated_at')
    }}
     - user
     p < likes
      - user
    p < likes
      - user
    x tags
`;


/*
* If the names of the tables and columns does not follow the ASSOCIATE rules, you need to manually insert the correct names of the tables and columns.
* members
*  - member_id
* articles
*  - id
*  - writer_id
* comments
*  - id
*  - article_id
*  - writer_id
* likes
*  - parent_name
*  - parent_id
*  - member_id
* tags_articles
*  - article_id
*  - tag_name
* tags
*  - name
* */

const posts = await ASSOCIATE `
  posts ${{
    table: 'articles'
  }}
    - user ${{
      left_key: 'writer_id',
      key: 'member_id',
      table: 'members'
    }}
    < comments ${{
      key: 'article_id'
    }}
      - user ${{
        left_key: 'writer_id',
        key: 'member_id',
        table: 'members'
      }}
      p < likes ${{
        poly_type: { parent_name: 'comments' },
        key: 'parent_id'
      }}
    p < likes ${{
      poly_type: { parent_name: 'articles' },
      key: 'parent_id'
    }}
    x tags ${{
      left_key: 'id',
      left_xkey: 'article_id',
      xtable: 'tags_articles',
      xkey: 'tag_name',
      key: 'name'
    }}
`;

If you use VIEW in databases, it's much easier. Then, you don't need to insert all correct column and table names.

ROW_NUMBER + PARTITION (PostgreSQL)

You can set the row_number option to fetch only up to four comments each post. Internally use ROW_NUMBER and PARTITION.

ASSOCIATE `
  posts ${SQL `WHERE is_hidden = false ORDER BY id DESC LIMIT ${10}`}
    < comments ${{
      row_number: [4, SQL `id DESC`]
    }}
`

Hook

You can add virtual columns, sorting, filtering and etc by using Hook. When all the datas are gathered below “posts”, Hook is executed.

const users = await ASSOCIATE `
  users ${{hook: users => users.map(u =>
    Object.assign({}, u, { _popular: !!u._.posts.find(p => p._is_best) })
  )}}
    < posts ${{hook: posts => posts.map(
      p => Object.assign({}, p, { _is_best: p._.comments.length > 1 }))}}
      - user
      < comments
       - user
`;

users[0]._popular; // true
users[0]._.posts[0]._is_best; // true
users[0]._.posts[1]._is_best; // false

ASSOCIATE_MODULE

ASSOCIATE allows you to modularize options for reuse. ASSOCIATE_MODULE in the function to be passed to ASSOCIATE.

Post.rights = () => ASSOCIATE_MODULE `
  - user
    < comments ${{
      row_number: [4, SQL `id DESC`]
    }}
     - user
     p < likes
      - user
    p < likes
      - user
    x tags
`;

ASSOCIATE `
  posts ${SQL `WHERE is_hidden = false ORDER BY id DESC LIMIT ${10}`}
    ${Post.rights}
`;

Use currying to pass arguments.

Post.rights = (limit = 4) => () => ASSOCIATE_MODULE `
  - user
    < comments ${{
      row_number: [limit, SQL `id DESC`]
    }}
     - user
     p < likes
      - user
    p < likes
      - user
    x tags
`;

ASSOCIATE `
  posts ${SQL `WHERE is_hidden = false ORDER BY id DESC LIMIT ${10}`}
    ${Post.rights(6)}
`;

Transaction

const { PostgreSQL } = require('mql2');
const { CONNECT } = PostgreSQL;
const POOL = CONNECT({
  host: 'localhost',
  user: 'username',
  password: '1234',
  database: 'dbname',
  charset: 'utf8'
});
const { TRANSACTION } = POOL;
const { QUERY, COMMIT, ROLLBACK } = await TRANSACTION();

await QUERY `
  INSERT INTO posts ${VALUES(post)}
`;
await QUERY `
  UPDATE posts ${SET({ body: 'yo!', updated_at: new Date() })} WHERE id = ${post.id}
`;
await ROLLBACK();

DEBUG

MQL_DEBUG.LOG = true;
QUERY `SELECT ${"hi~"} as ho`;

// { text: 'SELECT $1 as ho', values: ['hi'] }