NPM for Node.js set manipulation
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


A utility wrapper for set operations using arrays of objects. Work with sets as if they were simple arrays!

Each set has a chain of ancestry which, at it's root, is a collection. All values are held in the collections, while the sets themselves are simply orderings across subsets of ids


First, require the module:

var Indexed = require('indexed-set');

Then create a collection:

var collection = new Indexed.Collection(<array>, [<keyName>]);

where keyName defaults to Mongo's '_id' or

var collection = new Indexed.Collection(<options>);

The available options are:

  1. datasource : the mongo adapter
  2. name : the name of the collection to load
  3. onLoad : executed when the collection load is complete

then to create a set:

new Indexed.Set(collection);

If I want to initialize the set from a mongo collection using the built-in mongo adapter it would look like this:

var MongoAdapter = require('indexed-set/mongo');
var collection = new Indexed.Collection({
    name : 'my_awesome_collection',
    datasource : new MongoAdapter(options.database);,
    onLoad: function(){
        var mySet = new Indexed.Set(collection);
        //do stuff here

Internal Data Access

  1. getByPosition(position) : retrieve the object by the position of the object within the set
  2. getById(id) : retrieve the object by it's primary key within the set
  3. setByIdFromString(id, idString) : substitue the object with id idString for the position of the passed id
  4. setByPositionFromString(position, idString) : substitue the object with id idString for the passed position
  5. setByIdFromObject(id, object) : substitue this object at the position of the object with the passed in id
  6. setByPositionFromObject(position, object) : substitue this object at the position provided

additionally you are able to call:


which wraps a proxy around the object so you can use an array-like syntax:

so given a set:

var collection = new Indexed.Collection([
], 'a');
var mySet = new Indexed.Set(collection);

I could say:


which resolves to:


as does


and, as another example:




Meaning you can either use strings to address by primary key or integers to address by position. Proxies FTW.

Array Syntax

The Set class implements a number of the standard array methods, for familiarity

  1. indexOf(object)
  2. forEach(iteratorFunction)
  3. push(object)
  4. pop()
  5. shift()
  6. slice(start, [stop]) : take a subset by index range

Set Functions

This is the meat of the library allowing you to manipulate and analyze datasets while still staying economical with memory and performance.

  1. and(set) : the union of all objects in both sets
  2. or(set) : all objects in either set
  3. xor(set) : all objects contained in one set, but not both
  4. not(set) : all objects not present in 'set'
  5. filter(filterFunction) : a filter passed across the set to determine if a given element is selected or not
  6. with(field, comparison, value) : a convenience method to generate a filter
  7. byGroup(field) : segment along a given field and bin objects by the value on that field. returns an object of arrays
  8. distinct([field]) : return all potential values for a given field, or if none is provided, provide all options for all fields, broken down by field.
  9. clone() : make a copy of the set (this is a relatively cheap operation, in comparison to many clone() functions)
  10. resume() : allow filters to execute as they are applied (started by default), also executes queued filters
  11. pause() : queue applied filters instead of immediately executing them.
  12. filterFunction() : return a function which is a composite of all filter functions on this set(can be passed to mongo as a selector function).
  13. hierarchy(hierarchies, [fields]) : the tree of dependencies by node based on the descending relationships (provided an array of arrays of fieldNames). This produces overlapping relationship trees based on the passed priority. I won't go into further detail because this is a CPU bound task. If you don't know how to compensate for that, avoid the feature, lest you have unresponsive services.
  14. aggregate(field, [aggregatorFn]) : aggregate all set members on a given field, default aggregation is just a = a + b.


Run the test harness at the project root with:



-Abbey Hawk Sparrow