Aggsy

Aggsy is a aggregation language/module for easy use in http query strings

Try it online at https://aggsyplay.reminyborg.com

Installation

You need npm installed:

$ npm install aggsy

Example

var aggsy = require('aggsy')

var cars = [
  { model: 'volvo', make: 'v50', km: 100 },
  { model: 'tesla', make: 's', km: 200 },
  { model: 'tesla', make: 's', km: 120 },
  { model: 'tesla', make: 'x', km: 10 }
]

aggsy('model(distance: _sum(km), reports: _count())', cars)

// Gives:
{
  tesla: { reports: 2, distance: 330 },
  volvo: { reports: 1, distance: 100 }
}


// You may also flatten the results
aggsy('model(distance: _sum(km), reports: _count())', cars, { flatten: true })

// Gives:
[
  { model: 'tesla', reports: 2, distance: 330 },
  { model: 'volvo', reports: 1, distance: 100 }
]

To aggregate on nested groups

aggsy('model(make(count: _count()), count: _count())', cars)

// Gives:
{
  tesla: {
    's': { 'count': 2 },
    'x': { 'count': 1 },
    'count': 3
  },
  volvo: {
    'v50': { 'count': 1 },
    'count': 1
  }
}

// Flatten
aggsy('model(make(count: _count()), 'make.count': _count())', cars, { flatten: true })

// Gives:
[
  { model: 'tesla' make: 's', 'make.count': 2, count: 3 },
  { model: 'tesla' make: 'x', 'make.count': 1, count: 3 },
  { model: 'volvo': make: 'v50', 'make.count': 1, count: 1 }
}

Aggsy(query[, data, options])

When run with an aggsy query and array of objects the aggregated results is returned.

When run with only an query will return an [aggregate function](#advanced use)

Following options are available:

• reducers - optional list of [custom reducers](#custom reducers)
• missing - (default: false) grouping name to put items where grouping property does not exits
• flatten - (default: false) flatten result (may also be set with _flatten(query))

Query language

Grouping

Given a structure

{ model: 'volvo', details: { make: 'v50' }, km: 100 }

To group on model use `model()``

{ volvo: [/* items with model: volvo */] }

To group on make use dot notation `details.make()``

{ v50: [/* items with details.make: v50 */] }

If no reducers or nested groups are defined within a group ex: model() all the items are returned in the groups

Options

Flatten

_flatten(query)

Sets the flatten value so you get flat results

Reducers

Sum

_sum(property) Int

Count

_count() Int

Min

_min(property) Int

Max

_max(property) Int

First

_first(property)

Last

_last(property)

Has

_has(property) Bool

Average / Mean

_avg(property) { value: 0, count: 0 }

Standard deviation

_stdev(property) { value: 0, variance: 0, average: 0, count: 0 }

Static String

_static(string) String

Naming reducers

If only supplied with _sum(km) the result would be { '_sum(km)': 100 }

Reducers can be named with the convention distance: _sum(km) the result then would be { 'distance': 100 }

Custom reducers

You can add a list of custom reducers to the aggsy options object.

A reducer function behaves like javascript reduce but only the previousValue and currentValue is supplied to the reducer function. If you want to define an initialValue it must be added as a function property.

function _myownsum (prev, curr) { return prev + curr }
_myownsum.initialValue = 0
var options = { reducers: { '_myowsum': _myowsum } }

aggsy('_myownsum(km)', data, options)
// or
var aggregate = aggsy('_myownsum(km)', options)

License

MIT