Some Light Abstraction for RethinkDB

The RethinkDB Node driver is already very simple to use, but there are methods that I wish I "just had" at the ready. Those are:

• Query, which returns an array
• First, Exists
• Save, which upserts a record
• Automatic table/db config
• DB Manipulation (Create/Drop)

So I created it. That's what we have here.

In addition, I love Meteor.js and thought I would try to support its synchronous nature as best I could. I know that Meteor uses Fibers, but they can cause problems if you're not completely in control of them.

So, SecondThought uses the deasync library to "virtually unblock" the synchronous calls. The thread won't be blocked when you use these sync methods, but the consecutive code calls will. It's expermental, at best.

Usage

Install using

npm install secondthought --save

To use this in your code, just configure what you need:

var DB = require("second-thought");
DB.connect({db : "test"}, function(err,db){

  //you now have access to all of your tables as properties on your db variable:
  //so, assume there's a table called "foo" in your db...
  db.foo.save({name : "Mike"}, function(err,newRecord){

    //output the generated ID
    console.log(newRecord.id);
  });

});

You can do this synchronously as well:

var SecondThought = require("second-thought");
var db = SecondThought.connectSync({db : "test"});
var newRecord = db.foo.saveSync({name : "Pete"});

You can intermix the sync/async stuff all you like. Each method you see below has an sync counterpart which you can use by attaching Sync to the method name.

Each table on your DB object is a full-blown RethinkDB table, so you can step outside the abstraction at any point:

db.openConnection(function(err,conn){

  //this is a ReQL query
  db.foo.eqJoin('bar_id', db.bar).run(conn, function(err,cursor){

    //run the joined action and do something interesting
    cursor.toArray(function(err,array){
      //use the array...

      //be sure to close the connection!
      conn.close();
    });

  });
});

In addition you can do all kinds of fun things, like...

//installation of the DB and tables
db.connect({db : "test"}, function(err, db){
  db.install(['foo', 'bar'], function(err,result){
    //tables should be installed now...
  });
});

//add a secondary index
db.connect({db : "test"}, function(err,db){

  db.foo.index("email", function(err, indexed){
    //indexed == true;
  });
});

Basic Queries

I've tried to keep the API light and simple - with just a bit of sugar to keep the repetetive stuff to a minimum:

db.connect({db : "test", function(err,db){

  db.foo.query({category : "beer"}, function(err,beers){
    //beers is an array, so have at it
  });
  
  db.foo.contains({field : "name", vals : ["Bob", "Jill"]}, function(err,res){
    //res is two records with the names Bob and Jill
  });
  
  db.foo.first({email : "[email protected]"}, function(err,rob){
    //hi Rob
  });

  db.foo.exists({name : "bill"}, function(err, exists){
    //exists will tell you if it's there
  });

  db.foo.destroy({id : "some-id"}, function(err,destroyed){
    //destroyed will be true if something was deleted
  });

  db.foo.destroyAll(function(err,destroyed){
    //destroyed is the count of records whacked
  });

  db.foo.updateOnly({name : "Stevie"}, "some-id", function(err,result){
    //save will do a full swap of the document, updateOnly will partially update
    //a document so you need to pass the id along
    //result will be true if an update happened
  });

});

Have a look at the tests to see a bit more

Wanna Help?

Just do me a favor and open a PR with some ideas and hopefully a test or two. Thanks!