koa-rethinkdb-pool

Middleware that manages RethinkDB connections in a Koa application.

It provides functions on the Koa request context that allocate a connection when necessary and manage closing that connection when the request completes.

Example Usage

During startup, initialize the package with the RethinkDB instance and the connection information.

Supplying the RethinkDB instance ensures that your project is control of the RethinkDB package version. The connection information is simply the object passed to RethinkDB's connect function.

var rethinkdb = require('rethinkdb');

var pool = require('koa-rethinkdb-pool');

app.use(pool({
  r: rethinkdb,
  connectOptions: {
    host: 'localhost',
    port: 28015,
    db: 'test'
    }
}));

This installs middleware that will add the following functions to the Koa context:

• connection() - Return a Koa connection.
• execute(query) - Execute a query and return the result.
• fetchone(query) - Execute a query and return the first result value or null.
• fetchall(query) - Execute a query and return an array of results.

var r = require('rethinkdb');

app.post('/api/get', function*() {
  // Return all user objects
  this.body = yield this.fetchall(r.table('users'));
});

app.post('/api/getone', function*() {
  // Return the user object with id 1 or null if it doesn't exist.
  this.body = yield this.fetchone(r.table('users').get(1));
});

app.post('/api/insert', function*() {
  var result = yield this.execute(r.table('users').insert({ id: 7, name: 'seven' });
});

In each of these cases, a connection was allocated and attached to the request context (this). After the request completes, the middleware detects the connection and closes it.

Since the connection is attached to the request context, multiple calls will use the same connection:

app.post('/api/test', function*() {

  // This first call to execute makes the connection
  var result = yield this.execute(r.table('users').insert({ id: 7, name: 'seven' });

  // This will use the same connection.
  var row = yield this.execute(r.table('users').get(7));
});

API

package initialization

Requiring the package returns a setup function. The function accepts a single options hash that must contain:

• r - The RethinkDB instance to use.
• connectOptions - A JSON object passed to RethinkDB's connect when a connection is needed.

connection()

Returns the context's connection, allocating it if necessary. Subsequent calls within the same request always return the same connection.

execute(query)

Executes a query (which can be an insert or an update) and returns its result.

This is effectively return yield query.run(cnxn). The result is whatever `run returns.

fetchone(query)

Executes a query and returns the first result object or null if there are no objects.

This is designed for "select"-type queries that you expect to return a single object. It is a convenience wrapper that extracts the first element from the results and returns it. It returns null if the query did not return a result.

This method is designed for queries that return a cursor, but if the result is not a cursor, the result is returned directly.

fetchall(query)

Executes a query and returns the results as an array. If the query does not return results, an empty array is returned.

This method is designed for queries that return a cursor, but if the result is not a cursor, the result is returned directly. However, if the query does not return a cursor you should probably use execute instead for clarity.

Pooling

This package will eventually provide a connection pool, but the basic API here will not change. At this time, a new connection is created when necessary.