node package manager


A json document database built on node.js and leveldb


A json document database built on node.js and leveldb

This database is experimental, not for production use


This is a learning exercise. I wanted to combine LevelDB (using the levelup module)
with express to provide an HTTP interface and odata-parser to provide a querying syntax.


Neutrino is a web server, you read/write the data over http using either raw http requests, or the client library.

Install and start the Server

$ git clone
$ cd neutrinodb/database
$ [sudo] npm install
$ node server

Connect using the client

Install the neutrino package.

$ npm install neutrinodb

In your node application you can create a database client like this:

var client = require('neutrino')('http://localhost:8080/');

Create, list and delete tables

Neutrino has the concept of tables, you can create, list and delete them like this.

// create a new table called 'table1'
client.createTable('table1', function(err){
    if (err) console.log(err);
// list all tables in the database
client.tables(function(error, tables){
    if (err) console.log(err);
// delete 'table2' from the database. This will delete all data in the table
client.deleteTable('table2', function(err){
    if (err) console.log(err);

Create, get, list and delete objects

You can add, remove and list objects in your table like this:

// create an object withe the 'A' key 
client.put('test1', 'A', {foo:"bar"}, function(err){
    if (err) console.log(err);
// get object 'A' back
client.get('test1', 'A', function(err, data){
    if (err) console.log(err);
// delete object 'A'
client.del('test1', 'A', function(err){
    if (err) console.log(err);
// retrieve all records from table 1
client.query('table1', function(err, objects){

Querying tables

Tables can be queried using the OData syntax:

// all objects where foo = 'bar'.
client.query('table1', "foo eq 'bar'", function(err, data){ ... });
// first 5 objects matching the query
client.query('table1', "foo eq 'bar'", {top: 5}, function(err, data){ ... });
// second page of 5 objects matching the query
client.query('table1', "foo eq 'bar'", {top: 5, skip: 5}, function(err, data){ ... });

Note that querying a table like this will result in a full table scan.

If you only want one (or a few) records, use the 'top' option to improve performance.


To avoid full table scans, you can register 'views' which will be indexed for you by the database.

client.registerView('view1', 'table1', "foo eq 'bar'", function(err){ ... });
client.queryView('view1', function(err, data){ ... });

Views can also be parameterised, on the condition that only one parameter is used, and that's used to test if a property is 'eq'.

client.registerView('view1', 'table1', "foo eq '?'", function(err){ ... });
client.queryView('view1', 'bar', function(err, data){ ... });
client.queryView('view1', 'baz', function(err, data){ ... });

At the moment views cannot be deleted.