Our version of mysql that makes connecting to mysql simpler and more elegant. Especially made for our awesome expressjs boilerplate.
This module inherits the popular node-mysql.
npm install anytv-node-mysql --save
- Add a key-config pair once and use it anywhere
- Chain everything
- Do transactions elegantly
- Automatic rollback on transactions
- Easy to specify if you want a pool connection or a per-query connection
- Specify errors that can be retried and set their max retries (does not work on transactions yet)
- Auto-reconnect
- Easy debugging
- Use
.args(...)
to help closures
On your index.js / server.js / app.js, register your database using a key.
import mysql from 'anytv-node-mysql';
mysql.add('my_db', {
host: 'localhost',
user: 'root',
password: '',
database: test
});
After registering a db key and config, you can now start querying.
mysql.query(
'SELECT name FROM users WHERE name = ?',
['name'],
callback
)
.end();
Automatically closes the connection and does not work on transactions.
mysql.build(
'SELECT name FROM users WHERE name = ?',
['name']
)
.promise()
.then()
.catch();
Doing a single query using squel
Works on transactions.
const query = squel.select()
.from('users')
.field('name')
.where('name = ?', 'name');
mysql.squel(query, callback)
.end();
Doing a single query using squel and promise
const query = squel.select()
.field('name')
.from('users')
.where('name = ?', 'name');
mysql.build(query)
.promise()
.then()
.catch();
Doing a single query using knex and promise
const query = knex
.select('name')
.from('users')
.where('name', 'name');
mysql.build(query)
.promise()
.then()
.catch();
Note: Single connections will only be created on mysql.query(...)
while pooled connections will be created on mysql.use('db1', config.DB, true)
.
You can chain queries and use a single connection.
mysql.query(
'SELECT name FROM users WHERE name = ?',
['name'],
callback
)
.query(
'SELECT address FROM users WHERE name = ?',
['name2'],
callback2
)
.end();
In case you registered multiple databases, you can switch the current db by invoking mysql.use(key)
.
import mysql from 'anytv-node-mysql';
mysql.add('db_1', config.DB1);
mysql.add('db_2', config.DB2);
mysql.use('db1')
.query(
'SELECT name FROM users WHERE name = ?',
['name'],
callback
)
.end();
mysql.use('db2')
.query(
'SELECT name FROM users WHERE name = ?',
['name2'],
callback2
)
.end();
You can also switch databases right after ending.
mysql.use('db1')
.query(
'SELECT name FROM users WHERE name = ?',
['name'],
callback
)
.end() // if you forgot to call this, the .use('db2') will do it for you
.use('db2')
.query(
'SELECT name FROM users WHERE name = ?',
['name2'],
callback2
)
.end();
Creating a transaction returns a transaction object, not the mysql object. It's not possible to do another query after doing a commit.
mysql.use('db1')
.transaction()
.query('INSERT INTO unique_email(email) VALUES (?)', ['unique'], callback)
.query('INSERT INTO unique_email(email) VALUES (?)', ['unique'], callback)
.query('SELECT * FROM unique_email WHERE email = ?', ['unique'], callback)
.commit(final_callback);
If a query fails, the next queries won't be executed anymore. Plus, the callback
and final_callback
will have the same arguments.
Same with you, we had trouble with closures when we're querying inside a loop. We created a function where you can pass anything and we'll include it on the callback.
mysql.use('db1')
.args(obj, 2, 'config')
.query('INSERT INTO unique_email(email) VALUES (?)', ['unique'], callback)
.end();
function callback (err, result, args, last_query) {
args[0] === obj;
args[1] === 2;
args[2] === 'config';
}
We follow the node convention for callbacks. The error first followed by result.
mysql.use('db1')
.query('INSERT INTO unique_email(email) VALUES (?)', ['unique'], callback)
.end();
function callback (err, result, args, last_query) {
if (err) {
// do something with the error
...
return;
}
// do something with the result
...
}
The last executed query is accessible on the last_query
variable.
To solve intermittent issues, we added a function that accepts an array of error codes where the library will keep on retrying until it works or until it reaches the maximum retries. Default maximum retries is 3.
// Retryable errors from `.retry_if()` will be cleared on the next `mysql.use(key)`
mysql.use('db1')
.retry_if(['ER_LOCK_DEADLOCK', 'PROTOCOL_SEQUENCE_TIMEOUT'])
.set_max_retry(5)
.query('INSERT INTO unique_email(email) VALUES (?)', ['unique'], callback)
.end();
// Can also be set in the config and it will be applied on all queries on that database
mysql.add('db', {
host: 'localhost',
user: 'my_user',
password: 'my_password',
retryable_errors: ['ER_LOCK_DEADLOCK', 'PROTOCOL_SEQUENCE_TIMEOUT']
});
We added a .open
function that accepts a config.
mysql.open({
host: 'localhost',
user: dynamic_variable,
password: dynamic_variable2,
database: 'test'
})
.query('INSERT INTO unique_email(email) VALUES (?)', ['unique'], callback)
.end();
The mysql.use(...)
accepts a boolean 3rd parameter. If true, we'll use a pool connection and create it immediately. Calling .end()
on pooled connections won't do anything.
mysql.add('db1', config.DB1, true);
mysql.use('db1')
.query('INSERT INTO unique_email(email) VALUES (?)', ['unique'], callback)
.end();
Use mysql.end_pool() to close a pooled connection.
mysql.add('db1', config.DB1, true);
let db1 = mysql.use('db1');
db1.query('INSERT INTO unique_email(email) VALUES (?)', ['unique'], callback)
.end();
db1.query('INSERT INTO unique_email(email) VALUES (?)', ['uniqu3'], callback)
.end();
db1.end_pool();
mysql.set_logger(my_new_logger);
Code documentation can be found at here.
Install the tools needed:
npm install --only=dev
To compile the ES6 source code to ES5:
npm run build
npm run build -- --watch # for watching
To generate the docs:
npm run docs
Make sure to build the source code before running tests.
npm test
npm run coverage
Then open coverage/lcov-report/index.html.
MIT