budiadiono / mongoose-sequence

Sequence and autoincrement handling for mongoose

Mongoose sequence plugin

This plugin let you create fields which autoincrement their value:

• every time a new document is inserted in a collection
  or
• when you esplicitly want to increment them

This increment can be:

• global: every document has a unique value for the sequence field
• scoped: the counter depends on the value of other field(s)

Multiple counter can be set for a collection

Global sequences

Let's say you want to have an id field in your user collection which has an unique auto-incremented value.

The user schema is something like this:

UserSchema = mongoose.Schema({
    name: String
});

mongoose.model('User', UserSchema);

You don't need to define the id field in your schema because the plugin automatically set it for you. The only thing you have to do is to call:

UserSchema.plugin(AutoIncrement, {inc_field: 'id'});

Every time a new user is created, the id field will have an incremental number. The operation is atomic and is based on this specification. A commodity collection named counters is created for you. You can override the name of this collection but we will see this later with the options.

If you want to increment the _id field which is special to mongoose, you have to explicitly specify it as a Number and tell mongoose to not interfer:

UserSchema = mongoose.Schema({
    _id: Number,
    name: String
}, { _id: false });
UserSchema.plugin(AutoIncrement);

In this case you don't have to specify inc_field because the default value is _id

Not automatic sequences

Let say our user model has a like field which counts how many likes the user has in our amazing application. This field is of course a sequence but has to be incremented everytime an event occours. Because we have concurrent access to database we want to be sure that the increment of this counter happens safely. Let's start modify our schema

UserSchema = mongoose.Schema({
    name: String,
    like: Number
});

this time we specified explicitly the field like. There is no difference between defining and omitting the field specification. The only constraints is that the field has to be of type Number, otherwise the plugin will raise an error. So, let's say to the plugin we want the like field to be a safe counter

UserSchema.plugin(AutoIncrement, {inc_field: 'like', disable_hooks: true});

We specified disable_hooks. This avoid the field to be incremented when a new document is saved. So, how to increment this field? Your models has a new method: setNext. You must specify which sequence you want to increment and a callback. Here an example:

User.findOne({name:'George'}, function(err, user){
    user.setNext('like', function(err, user){
        if(err) console.log('Cannot increment the likes because',err);
    });
});

You noticed that the method setNext takes, as argument, the counter field name. Is possible to give a name to the counter and use it as reference. For the previous example we can define the counter like this:

UserSchema.plugin(AutoIncrement, {id:'like_counter', inc_field: 'like', disable_hooks: true});

and then using

user.setNext('like_counter', function(err, user){
    ...
});

So, if you not specify the id, the field name is used. Even if you're not forced to specify an id, its use is strongly suggested. This because if you have two different counters, which refers to fields with the same name, they will collide and incrementing one, will increment the other too. So use unique id to be sure to avoid collision.

As we will see, the use of an id for the counter is mandatory when you're are defining a scoped counter.

Advanced

Scoped counters

Let say our users are organized for country and city. And we want to save the inhabitant_number according to the two informations.
The schema is like this:

UserSchema = mongoose.Schema({
    name: String,
    country: String,
    city: String,
    inhabitants: Number
});

Every time a new Parisian is added the counting of Parisians have to be incremented. The inhabitants of New York must not interfer and have their separated counting. We should define a scoped counter which says, increment the counter, depending on the value of other fields.

UserSchema.plugin(AutoIncrement, {id: 'inhabitant_seq', inc_field: 'inhabitants', reference_fields: ['country','city'] });

Notice that we have to use an id for our sequence, otherwise the plugin will raise an error. Now save a new user

var user = new User({
    name: 'Patrice',
    country: 'France',
    city: 'Paris'
});
user.save();

This user will have the inhabitants counter to 1. If now we add a new inhabitant from New York, this will have the counter to 1 also, because the counter is referenced to the value of the fields country and city.

If we want to increment manually this counter we have to specify the id of the sequence in the setNext method

user.setNext('inhabitant_seq', function(err, user){
    user.inhabitants; // the counter value
});

Of course this example is a bit forced and this is for sure not the perfect use case. The field country and city have to be present and must not change during the life of the document because no automatic hook are set on the change of those values. But there are situation when you want a similar behaviour.

Options

This plugin accept a series of options.

• inc_field: The name of the field to increment. Mandatory, default is _id
• id: Id of the sequence. Is mandatory only for scoped sequences but its use is strongly encouraged.
• reference_fields: The field to reference for a scoped counter. Optional
• disable_hooks: If true, the counter will not be incremented on saving a new document
• collection_name: By default the collection name for the counters is counters. You can override it using this option