Run results for: Temp obj or in check

Run details:

User agent: Mozilla/5.0 (iPhone; CPU iPhone OS 10_1 like Mac OS X) AppleWebKit/602.1.32 (KHTML, like Gecko) Mobile/14B72c Twitter for iPhone

Browser: Mobile Safari 10

Operating system: iOS 10.1

Device Platform: iPhone

Date tested: 7 years ago

Test name	Executions per second
Empty object with short method	1139251.0 Ops/sec
Empty object with long method	1409699.0 Ops/sec
Full object with short method	1139789.0 Ops/sec
Full object with long method	1031442.0 Ops/sec
Empty object with shorter method	1396127.0 Ops/sec
Full object with shorter method	1119862.0 Ops/sec

Script Preparation code:

var haystack = {};
var SIZE = 10000;
for (var i = 0; i < SIZE; i++) { haystack['a' + i] = i; }
var objA = {};
var objB = { haystack: haystack };
function get1(obj, key) { return (obj.haystack || {})[key] || -1; }
function get2(obj, key) { return (obj.haystack && key in obj.haystack) ? obj.haystack[key] : -1; }
function get3(obj, key) { return obj.haystack && obj.haystack[key] || -1; }
function randomKey() { return 'a' + Math.floor(Math.random() * SIZE); }

Tests:

Empty object with short method
get1(objA, randomKey());
Empty object with long method
get2(objA, randomKey());
Full object with short method
get1(objB, randomKey());
Full object with long method
get2(objB, randomKey());
Empty object with shorter method
get3(objA, randomKey());
Full object with shorter method
get3(objB, randomKey());

Run results for: Temp obj or in check

Is it faster to test if an object contains a sub-object before looking up keys on it, or substituting an empty object in a falsy check

Empty object with short method

Empty object with long method

Full object with short method

Full object with long method

Empty object with shorter method

Full object with shorter method