groupByAsync

Group values according to an indicator function.

Usage

var groupByAsync = require( '@stdlib/utils/async/group-by' );

groupByAsync( collection, [options,] indicator, done )

Groups values according to an indicator function, which specifies which group an element in the input collection belongs to.

function indicator( value, next ) {
    setTimeout( onTimeout, value );
    function onTimeout() {
        console.log( value );
        next( null, (value > 2000) );
    }
}

function done( error, result ) {
    if ( error ) {
        throw error;
    }
    console.log( result );
}

var arr = [ 3000, 2500, 1000 ];

groupByAsync( arr, indicator, done );
/* =>
    1000
    2500
    3000
    { 'true': [ 2500, 3000 ], 'false': [ 1000 ] }
*/

The function accepts the following options:

limit: the maximum number of pending invocations at any one time. Default: infinity.
series: boolean indicating whether to sequentially invoke the indicator function for each collection element. If true, the function sets options.limit=1. Default: false.
returns: specifies the output format. If the option equals 'values', the function returns group results as element values. If the option equals 'indices', the function returns group results as element indices. If the option equals '*', the function returns group results as both element indices and values. Default: 'values'.
thisArg: the execution context for indicator.

By default, the function returns group results as element values. To return element indices, set the returns option to 'indices'.

function indicator( value, next ) {
    setTimeout( onTimeout, value );
    function onTimeout() {
        console.log( value );
        next( null, (value > 2000) );
    }
}

function done( error, result ) {
    if ( error ) {
        throw error;
    }
    console.log( result );
}

var arr = [ 3000, 2500, 1000 ];

var opts = {
    'returns': 'indices'
};
groupByAsync( arr, opts, indicator, done );
/* =>
    1000
    2500
    3000
    { 'true': [ 1, 0 ], 'false': [ 2 ] }
*/

To return index-value pairs, set the returns option to '*'.

function indicator( value, next ) {
    setTimeout( onTimeout, value );
    function onTimeout() {
        console.log( value );
        next( null, (value > 2000) );
    }
}

function done( error, result ) {
    if ( error ) {
        throw error;
    }
    console.log( result );
}

var arr = [ 3000, 2500, 1000 ];

var opts = {
    'returns': '*'
};
groupByAsync( arr, opts, indicator, done );
/* =>
    1000
    2500
    3000
    { 'true': [ [ 1, 2500 ], [ 0, 3000 ] ], 'false': [ [ 2, 1000 ] ] }
*/

By default, all elements are processed concurrently, which means that the function does not guarantee completion order. To process each collection element sequentially, set the series option to true.

function indicator( value, next ) {
    setTimeout( onTimeout, value );
    function onTimeout() {
        console.log( value );
        next( null, (value > 2000) );
    }
}

function done( error, result ) {
    if ( error ) {
        throw error;
    }
    console.log( result );
}

var arr = [ 3000, 2500, 1000 ];

var opts = {
    'series': true
};

groupByAsync( arr, opts, indicator, done );
/* =>
    3000
    2500
    1000
    { 'true': [ 3000, 2500 ], 'false': [ 1000 ] }
*/

To limit the maximum number of pending function invocations, set the limit option.

function indicator( value, next ) {
    setTimeout( onTimeout, value );
    function onTimeout() {
        console.log( value );
        next( null, (value > 2000) );
    }
}

function done( error, result ) {
    if ( error ) {
        throw error;
    }
    console.log( result );
}

var arr = [ 3000, 2500, 1000 ];

var opts = {
    'limit': 2
};

groupByAsync( arr, opts, indicator, done );
/* =>
    2500
    3000
    1000
    { 'true': [ 2500, 3000 ], 'false': [ 1000 ] }
*/

To set the execution context of the indicator function, set the thisArg option.

function indicator( value, next ) {
    this.count += 1;
    setTimeout( onTimeout, value );
    function onTimeout() {
        next( null, (value > 2000) );
    }
}

var arr = [ 3000, 2500, 1000 ];

var context = {
    'count': 0
};

var opts = {
    'thisArg': context
};

groupByAsync( arr, opts, indicator, done );

function done( error, result ) {
    if ( error ) {
        throw error;
    }
    console.log( result );
    // => { 'true': [ 2500, 3000 ], 'false': [ 1000 ] }

    console.log( context.count );
    // => 3
}

When invoked, the indicator function is provided a maximum of four arguments:

value: collection value.
index: collection index.
collection: the input collection.
next: a callback which should be called once the indicator function has finished processing a collection value.

The actual number of provided arguments depends on function length. If the indicator function accepts two arguments, the indicator function is provided value and next. If the indicator function accepts three arguments, the indicator function is provided value, index, and next. For every other indicator function signature, the indicator function is provided all four arguments.

function indicator( value, i, collection, next ) {
    console.log( 'collection: %s. %d: %d', collection.join( ',' ), i, value );
    setTimeout( onTimeout, value );
    function onTimeout() {
        console.log( value );
        next( null, (value > 2000) );
    }
}

function done( error, result ) {
    if ( error ) {
        throw error;
    }
    console.log( result );
}

var arr = [ 3000, 2500, 1000 ];

groupByAsync( arr, indicator, done );
/* =>
    collection: 3000,2500,1000. 0: 3000
    collection: 3000,2500,1000. 1: 2500
    collection: 3000,2500,1000. 2: 1000
    1000
    2500
    3000
    { 'true': [ 2500, 3000 ], 'false': [ 1000 ] }
*/

groupByAsync.factory( [options,] indicator )

Returns a function which invokes an indicator function once for each element in a collection.

function indicator( value, next ) {
    setTimeout( onTimeout, value );
    function onTimeout() {
        console.log( value );
        next( null, (value > 2000) );
    }
}

function done( error, result ) {
    if ( error ) {
        throw error;
    }
    console.log( result );
}

var f = groupByAsync.factory( indicator );

var arr1 = [ 3000, 2500, 1000 ];

f( arr1, done );
/* =>
    1000
    2500
    3000
    { 'true': [ 2500, 3000 ], 'false': [ 1000 ] }
*/

var arr2 = [ 300, 250, 100 ];

f( arr2, done );
/* =>
    100
    250
    300
    { 'false': [ 100, 250, 300 ] }
*/

The function accepts the same options as groupByAsync().

Notes

A collection may be either an Array, Typed Array, or an array-like Object (excluding strings and functions).
If a provided function calls the next callback with a truthy error argument, the function suspends execution and immediately calls the done callback for subsequent error handling.
The function does not support dynamic collection resizing.
The function does not skip undefined elements.
If provided an empty collection, the function calls the done callback with an empty object for the group results.
Neither groupByAsync nor the function returned by the factory method guarantee asynchronous execution. To guarantee asynchrony, wrap the done callback in a function which either executes at the end of the current stack (e.g., nextTick) or during a subsequent turn of the event loop (e.g., setImmediate, setTimeout).

The group returned by an indicator function should be a value which can be serialized as an object key. As a counterexample,

function indicator( value, next ) {
    setTimeout( onTimeout, value );
    function onTimeout() {
        console.log( value );
        next( null, {} );
    }
}

function done( error, result ) {
    if ( error ) {
        throw error;
    }
    console.log( result );
}

var arr = [ 3000, 2500, 1000 ];

groupByAsync( arr, indicator, done );
/* =>
    1000
    2500
    3000
    { '[object Object]': [ 1000, 2500, 3000 ] }
*/

while each group identifier is unique, all collection elements resolve to the same group because each group identifier serializes to the same string.

Examples

var resolve = require( 'path' ).resolve;
var readFile = require( '@stdlib/fs/read-file' );
var groupByAsync = require( '@stdlib/utils/async/group-by' );

var files = [
    resolve( __dirname, 'package.json' ),
    resolve( __dirname, 'README.md' ),
    resolve( __dirname, 'beep.boop.md' )
];

function done( error, result ) {
    if ( error ) {
        throw error;
    }
    console.log( result );
}

function indicator( file, next ) {
    var opts = {
        'encoding': 'utf8'
    };
    readFile( file, opts, onFile );

    function onFile( error ) {
        if ( error ) {
            return next( null, 'nonreadable' );
        }
        next( null, 'readable' );
    }
}

groupByAsync( files, indicator, done );