JavaScript: weird boolean values for [] - javascript

I've been learning JavaScript for some time now and I know better than to actually rely on this kind of behavior but I'm getting curious... why does the empty array [] have seemingly conflicting boolean identities?
It is truthy: [] && true evaluates to true ...and yet, it is false: [] == false evaluates to true.
However, its double negation is true: !![] evaluates to true. But it's also (using loose equality) equal to its own negation: ![] == [] evaluates to true.
Its numeric value is falsy: +[] && true evaluates to 0.
Its string representation is falsy: [].toString() && true evaluates to "".
Is there any clear reason why [] is such a conundrum?


I think the results are reasonable. The only one I'm iffy on is +[], but I might have that one figured out. It's like you said, loose comparison can be misleading, but it makes sense to me.
console.log([] && true); //true - there is an array
console.log([] == false); //true, the array is empty
console.log(![]); //false, there is an array
console.log(!![]); //true, there isn't not an array
console.log(![] == []); //true, no array is loosely like an empty array
console.log(+[]); //0, I think this converts the array to a number, 0 being the logical choice. +[1] will return "1" and +[1,2] is "NaN" - which seems to be because the conversion is no longer feasible.
console.log([].toString()); //string '', as expected - also see next example
console.log(['stuff',{}].toString()); //stuff,[object Object]


I can't give you a full answer, but here are some observations:
As you know there's a Boolean function that returns a boolean value and it accepts a parameter, thus it closely resembles casting to bool. Boolean([]) returns true which explains why [] && true is true, ![] is false and !![] is true.
On the other hand the Number([]) constructor returns 0, which explains why +[] && true is zero.
Another thing is that String([]) is empty string, String(true) is the string "true" and "" && "true" is "" which perfectly fits into my grand scheme as you'll see below.
Now, about the curious things. Here's what I found out:
Number(![]) is 0, Number([]) is 0 as well. In that case Number(![]) == Number([]) is obviously true.
Number([]) is 0, Number(false) is 0, thus Number([]) == Number(false).
It seems to me that whenever JS can't directly compare two objects (neither of which are strings) it casts them to numbers and then compares. If one of them is a string it always converts to string and then evaluates any functions. At least that would explain the strange behaviour you've noticed.


&& Returns expr1 if it can be converted to false; otherwise,
returns expr2. Thus, when used with Boolean values, && returns true if
both operands can be converted to true; otherwise, returns false.
https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Logical_Operators
In your case
+[] // 0 - is falsy so it will be returned
[].toString() // "" - is falsy so it will be returned

Related

What is wrong in this code, why does it print only the first code block [duplicate]

What are the values in JavaScript that are 'falsey', meaning that they evaluate as false in expressions like if(value), value ? and !value?
There are some discussions of the purpose of falsey values on Stack Overflow already, but no exhaustive complete answer listing what all the falsey values are.
I couldn't find any complete list on MDN JavaScript Reference, and I was surprised to find that the top results when looking for a complete, authoritative list of falsey values in JavaScript were blog articles, some of which had obvious omissions (for example, NaN), and none of which had a format like Stack Overflow's where comments or alternative answers could be added to point out quirks, surprises, omissions, mistakes or caveats. So, it seemed to make sense to make one.

Falsey values in JavaScript
false
Zero of Number type: 0 and also -0, 0.0, and hex form 0x0 (thanks RBT)
Zero of BigInt type: 0n and 0x0n (new in 2020, thanks GetMeARemoteJob)
"", '' and `` - strings of length 0
null
undefined
NaN
document.all (in HTML browsers only)
This is a weird one. document.all is a falsey object, with typeof as undefined. It was a Microsoft-proprietory function in IE before IE11, and was added to the HTML spec as a "willful violation of the JavaScript specification" so that sites written for IE wouldn't break on trying to access, for example, document.all.something; it's falsy because if (document.all) used to be a popular way to detect IE, before conditional comments. See Why is document.all falsy? for details
"Falsey" simply means that JavaScript's internal ToBoolean function returns false. ToBoolean underlies !value, value ? ... : ...; and if (value). Here's its official specification (2020 working draft) (the only changes since the very first ECMAscript specification in 1997 are the addition of ES6's Symbols, which are always truthy, and BigInt, mentioned above:
Argument type
Result
Undefined
Return false.
Null
Return false.
Boolean
Return argument.
Number
If argument is +0, -0, or NaN, return false; otherwise return true.
String
If argument is the empty String (its length is zero), return false; otherwise return true.
BigInt
If argument is 0n, return false; otherwise return true.
Symbol
Return true.
Object
Return true.
Comparisons with == (loose equality)
It's worth talking about falsy values' loose comparisons with ==, which uses ToNumber() and can cause some confusion due to the underlying differences. They effectively form three groups:
false, 0, -0, "", '' all match each other with ==
e.g. false == "", '' == 0 and therefore 4/2 - 2 == 'some string'.slice(11);
null, undefined match with ==
e.g. null == undefined but undefined != false
It's also worth mentioning that while typeof null returns 'object', null is not an object, this is a longstanding bug/quirk that was not fixed in order to maintain compatibility. It's not a true object, and objects are truthy (except for that "wilful violation" document.all when Javascript is implemented in HTML)
NaN doesn't match anything, with == or ===, not even itself
e.g. NaN != NaN, NaN !== NaN, NaN != false, NaN != null
With "strict equality" (===), there are no such groupings. Only false === false.
This is one of the reasons why many developers and many style guides (e.g. standardjs) prefer === and almost never use ==.
Truthy values that actually == false
"Truthy" simply means that JavaScript's internal ToBoolean function returns true. A quirk of Javascript to be aware of (and another good reason to prefer === over ==): it is possible for a value to be truthy (ToBoolean returns true), but also == false.
You might think if (value && value == false) alert('Huh?') is a logical impossibility that couldn't happen, but it will, for:
"0" and '0' - they're non-empty strings, which are truthy, but Javascript's == matches numbers with equivalent strings (e.g. 42 == "42"). Since 0 == false, if "0" == 0, "0" == false.
new Number(0) and new Boolean(false) - they're objects, which are truthy, but == sees their values, which == false.
0 .toExponential(); - an object with a numerical value equivalent to 0
Any similar constructions that give you a false-equaling value wrapped in a type that is truthy
[], [[]] and [0] (thanks cloudfeet for the JavaScript Equality Table link)
Some more truthy values
These are just a few values that some people might expect to be falsey, but are actually truthy.
-1 and all non-zero negative numbers
' ', " ", "false", 'null'... all non-empty strings, including strings that are just whitespace
Anything from typeof, which always returns a non-empty string, for example:
typeof null (returns a string 'object' due to a longstanding bug/quirk)
typeof undefined (returns a string 'undefined')
Any object (except that "wilful violation" document.all in browsers). Remember that null isn't really an object, despite typeof suggesting otherwise. Examples:
{}
[]
function(){} or () => {} (any function, including empty functions)
Error and any instance of Error
Any regular expression
Anything created with new (including new Number(0) and new Boolean(false))
Any Symbol
true, 1, "1" and [1] return true when compared to each other with ==.

Don't forget about the non-empty string "false" which evaluates to true

Just to add to #user568458's list of falsy values:
In addition to integer number 0, the decimal number 0.0, 0.00 or any such zeroish number is also a falsy value.
var myNum = 0.0;
if(myNum){
console.log('I am a truthy value');
}
else {
console.log('I am a falsy value');
}
Above code snippet prints I am a falsy value
Similarly hex representation of the number 0 is also a falsy value as shown in below code snippet:
var myNum = 0x0; //hex representation of 0
if(myNum){
console.log('I am a truthy value');
}
else {
console.log('I am a falsy value');
}
Above code snippet again prints I am a falsy value.

Addition to the topic, as of ES2020 we have a new value which is falsy, it's BigInt zero (0n):
0n == false // true
-0n == false // true
0n === false // false
-0n === false // false
So with this, we now have 7 "falsy" values in total (not including document.all as mentioned by user above since it's part of DOM and not JS).

What is the different when we use !! and no use it in condition expression? [duplicate]

I saw some code that seems to use an operator I don't recognize, in the form of two exclamation points, like so: !!. Can someone please tell me what this operator does?
The context in which I saw this was,
this.vertical = vertical !== undefined ? !!vertical : this.vertical;

It converts Object to boolean. If it was falsey (e.g., 0, null, undefined, etc.), it would be false, otherwise, true.
!object // Inverted Boolean
!!object // Noninverted Boolean, so true Boolean representation
So !! is not an operator; it's just the ! operator twice.
It may be simpler to do:
Boolean(object) // Boolean
Real World Example "Test IE version":
const isIE8 = !! navigator.userAgent.match(/MSIE 8.0/);
console.log(isIE8); // Returns true or false
If you ⇒
console.log(navigator.userAgent.match(/MSIE 8.0/));
// Returns either an Array or null
But if you ⇒
console.log(!!navigator.userAgent.match(/MSIE 8.0/));
// Returns either true or false

It's a horribly obscure way to do a type conversion.
! means NOT. So !true is false, and !false is true. !0 is true, and !1 is false.
So you're converting a value to a Boolean, inverting it, and then inverting it again.
// Maximum Obscurity:
val.enabled = !!userId;
// Partial Obscurity:
val.enabled = (userId != 0) ? true : false;
// And finally, much easier to understand:
val.enabled = (userId != 0);
// Or just
val.enabled = Boolean(userId);
Note: the latter two expressions aren't exactly equivalent to the first expression when it comes to some edge case (when userId is [], for example), due to the way the != operator works and what values are considered truthy.

!!expr (two ! operators followed by an expression) returns a Boolean value (true or false) depending on the truthiness of the expression. It makes more sense when used on non-boolean types. Consider these examples, especially the 3rd example and onward:
!!false === false
!!true === true
!!0 === false
!!parseInt("foo") === false // NaN is falsy
!!1 === true
!!-1 === true // -1 is truthy
!!(1/0) === true // Infinity is truthy
!!"" === false // empty string is falsy
!!"foo" === true // non-empty string is truthy
!!"false" === true // ...even if it contains a falsy value
!!window.foo === false // undefined value is falsy
!!undefined === false // undefined primitive is falsy
!!null === false // null is falsy
!!{} === true // an (empty) object is truthy
!![] === true // an (empty) array is truthy; PHP programmers beware!

Brew some tea:
!! is not an operator. It is the double-use of ! -- which is the logical "not" operator.
In theory:
! determines the "truth" of what a value is not:
The truth is that false is not true (that's why !false results
in true)
The truth is that true is not false (that's why !true results
in false)
!! determines the "truth" of what a value is not not:
The truth is that true is not not true (that's why !!true results in true)
The truth is that false is not not false (that's why !!false results in false)
What we wish to determine in the comparison is the "truth" about the value of a reference, not the value of the reference itself. There is a use-case where we might want to know the truth about a value, even if we expect the value to be false (or falsey), or if we expect the value not to be typeof boolean.
In practice:
Consider a concise function which detects feature functionality (and in this case, platform compatibility) by way of dynamic typing (aka "duck typing"). We want to write a function that returns true if a user's browser supports the HTML5 <audio> element, but we don't want the function to throw an error if <audio> is undefined; and we don't want to use try ... catch to handle any possible errors (because they're gross); and also we don't want to use a check inside the function that won't consistently reveal the truth about the feature (for example, document.createElement('audio') will still create an element called <audio> even if HTML5 <audio> is not supported).
Here are the three approaches:
// this won't tell us anything about HTML5 `<audio>` as a feature
var foo = function(tag, atr) { return document.createElement(tag)[atr]; }
// this won't return true if the feature is detected (although it works just fine)
var bar = function(tag, atr) { return !document.createElement(tag)[atr]; }
// this is the concise, feature-detecting solution we want
var baz = function(tag, atr) { return !!document.createElement(tag)[atr]; }
foo('audio', 'preload'); // returns "auto"
bar('audio', 'preload'); // returns false
baz('audio', 'preload'); // returns true
Each function accepts an argument for a <tag> and an attribute to look for, but they each return different values based on what the comparisons determine.
But wait, there's more!
Some of you probably noticed that in this specific example, one could simply check for a property using the slightly more performant means of checking if the object in question has a property. There are two ways to do this:
// the native `hasOwnProperty` method
var qux = function(tag, atr) { return document.createElement(tag).hasOwnProperty(atr); }
// the `in` operator
var quux = function(tag, atr) { return atr in document.createElement(tag); }
qux('audio', 'preload'); // returns true
quux('audio', 'preload'); // returns true
We digress...
However rare these situations may be, there may exist a few scenarios where the most concise, most performant, and thus most preferred means of getting true from a non-boolean, possibly undefined value is indeed by using !!. Hopefully this ridiculously clears it up.

!! converts the value to the right of it to its equivalent Boolean value. (Think poor man's way of "type-casting".) Its intent is usually to convey to the reader that the code does not care what value is in the variable, but what its "truth" value is.

!!foo applies the unary not operator twice and is used to cast to a Boolean type similar to the use of unary plus +foo to cast to a number and concatenating an empty string ''+foo to cast to a string.
Instead of these hacks, you can also use the constructor functions corresponding to the primitive types (without using new) to explicitly cast values, i.e.,
Boolean(foo) === !!foo
Number(foo) === +foo
String(foo) === ''+foo

So many answers doing half the work. Yes, !!X could be read as "the truthiness of X [represented as a Boolean]". But !! isn't, practically speaking, so important for figuring out whether a single variable is (or even if many variables are) truthy or falsy. !!myVar === true is the same as just myVar. Comparing !!X to a "real" Boolean isn't really useful.
The only thing you gain with !! is the ability to check the truthiness of multiple variables against each other in a repeatable, standardized (and JSLint friendly) fashion.
Simply casting :(
That is...
0 === false is false.
!!0 === false is true.
The above's not so useful. if (!0) gives you the same results as if (!!0 === false). I can't think of a good case for casting a variable to Boolean and then comparing to a "true" Boolean.
See "== and !=" from JSLint's directions (note: Crockford is moving his site around a bit; that link is liable to die at some point) for a little on why:
The == and != operators do type coercion before comparing. This is bad because it causes ' \t\r\n' == 0 to be true. This can mask type errors. JSLint cannot reliably determine if == is being used correctly, so it is best to not use == and != at all and to always use the more reliable === and !== operators instead.
If you only care that a value is truthy or falsy, then use the short form. Instead of
(foo != 0)
just say
(foo)
and instead of
(foo == 0)
say
(!foo)
Note that there are some unintuitive cases where a Boolean will be cast to a number (true is cast to 1 and false to 0) when comparing a Boolean to a number. In this case, !! might be mentally useful. Though, again, these are cases where you're comparing a non-Boolean to a hard-typed Boolean, which is, in my opinion, a serious mistake. if (-1) is still the way to go here.
╔═══════════════════════════════════════╦═══════════════════╦═══════════╗
║ Original ║ Equivalent ║ Result ║
╠═══════════════════════════════════════╬═══════════════════╬═══════════╣
║ if (-1 == true) console.log("spam") ║ if (-1 == 1) ║ undefined ║
║ if (-1 == false) console.log("spam") ║ if (-1 == 0) ║ undefined ║
║ Order doesn't matter... ║ ║ ║
║ if (true == -1) console.log("spam") ║ if (1 == -1) ║ undefined ║
╠═══════════════════════════════════════╬═══════════════════╬═══════════╣
║ if (!!-1 == true) console.log("spam") ║ if (true == true) ║ spam ║ better
╠═══════════════════════════════════════╬═══════════════════╬═══════════╣
║ if (-1) console.log("spam") ║ if (truthy) ║ spam ║ still best
╚═══════════════════════════════════════╩═══════════════════╩═══════════╝
And things get even crazier depending on your engine. WScript, for instance, wins the prize.
function test()
{
return (1 === 1);
}
WScript.echo(test());
Because of some historical Windows jive, that'll output -1 in a message box! Try it in a cmd.exe prompt and see! But WScript.echo(-1 == test()) still gives you 0, or WScript's false. Look away. It's hideous.
Comparing truthiness :)
But what if I have two values I need to check for equal truthiness/falsiness?
Pretend we have myVar1 = 0; and myVar2 = undefined;.
myVar1 === myVar2 is 0 === undefined and is obviously false.
!!myVar1 === !!myVar2 is !!0 === !!undefined and is true! Same truthiness! (In this case, both "have a truthiness of falsy".)
So the only place you'd really need to use "Boolean-cast variables" would be if you had a situation where you're checking if both variables have the same truthiness, right? That is, use !! if you need to see if two variables are both truthy or both falsy (or not), that is, of equal (or not) truthiness.
I can't think of a great, non-contrived use case for that offhand. Maybe you have "linked" fields in a form?
if (!!customerInput.spouseName !== !!customerInput.spouseAge ) {
errorObjects.spouse = "Please either enter a valid name AND age "
+ "for your spouse or leave all spouse fields blank.";
}
So now if you have a truthy for both or a falsy for both spouse name and age, you can continue. Otherwise you've only got one field with a value (or a very early arranged marriage) and need to create an extra error on your errorObjects collection.
Though even in this case, the !! really is superfluous. One ! was enough to cast to a Boolean, and you're just checking equality.
EDIT 24 Oct 2017, 6 Feb 19:
Third-party libraries that expect explicit Boolean values
Here's an interesting case... !! might be useful when third-party libraries expect explicit Boolean values.
React
For instance, False in JSX (React) has a special meaning that's not triggered on simple falsiness. If you tried returning something like the following in your JSX, expecting an int in messageCount...
{messageCount && <div>You have messages!</div>}
... you might be surprised to see React render a 0 when you have zero messages. You have to explicitly return false for JSX not to render. The above statement returns 0, which JSX happily renders, as it should. It can't tell you didn't have Count: {messageCount}.
One fix involves the bangbang, which coerces 0 into !!0, which is false:
{!!messageCount && <div>You have messages!</div>}
JSX' documentation suggests you be more explicit, write self-commenting code, and use a comparison to force to a Boolean.
{messageCount > 0 && <div>You have messages!</div>}
I'm more comfortable handling falsiness myself with a ternary --
{messageCount ? <div>You have messages!</div> : false}
TypeScript
The same deal in TypeScript: If you have a function that returns a Boolean (or you're assigning a value to a Boolean variable), you [usually] can't return/assign a boolean-y value; it has to be a strongly typed boolean. This means, iff myObject is strongly typed, return !myObject; works for a function returning a Boolean, but return myObject; doesn't. You have to return !!myObject (or cast to the proper Boolean another way) to match TypeScript's expectations.
The exception for TypeScript? If myObject was an any, you're back in JavaScript's Wild West and can return it without !!, even if your return type is a Boolean.
Keep in mind that these are JSX and TypeScript conventions, not ones inherent to JavaScript.
But if you see strange 0s in your rendered JSX, think loose falsy management.

It's just the logical NOT operator, twice. It's used to convert something to Boolean, e.g.:
true === !!10
false === !!0

It converts the suffix to a Boolean value.

It's a double not operation. The first ! converts the value to Boolean and inverts its logical value. The second ! inverts the logical value back.

It seems that the !! operator results in a double negation.
var foo = "Hello, World!";
!foo // Result: false
!!foo // Result: true

!! is using the NOT operation twice together. ! converts the value to a Boolean and reverses it, so using it twice, showing the Boolean (false or true) of that value. Here is a simple example to see how !! works:
At first, the place you have:
var zero = 0;
Then you do !0. It will be converted to Boolean and be evaluated to true, because 0 is falsy, so you get the reversed value and converted to Boolean, so it gets evaluated to true.
!zero; //true
But we don't want the reversed Boolean version of the value, so we can reverse it again to get our result! That's why we use another !.
Basically, !! makes us sure the value we get is Boolean, not falsy, truthy, string, etc...
So it's like using the Boolean function in JavaScript, but an easier and shorter way to convert a value to Boolean:
var zero = 0;
!!zero; //false

It simulates the behavior of the Boolean() casting function.
The first NOT returns a Boolean value no matter what operand it is given. The second NOT negates that Boolean value and so gives the true Boolean value of a variable. The end result is the same as using the Boolean() function on a value.

! is "Boolean not", which essentially typecasts the value of "enable" to its boolean opposite. The second ! flips this value. So, !!enable means "not not enable," giving you the value of enable as a Boolean.

I think worth mentioning is that a condition combined with logical AND/OR will not return a Boolean value, but the last success or first fail in case of && and the first success or last fail in case of || of the condition chain.
res = (1 && 2); // res is 2
res = (true && alert) // res is function alert()
res = ('foo' || alert) // res is 'foo'
In order to cast the condition to a true Boolean literal we can use the double negation:
res = !!(1 && 2); // res is true
res = !!(true && alert) // res is true
res = !!('foo' || alert) // res is true

The !! construct is a simple way of turning any JavaScript expression into
its Boolean equivalent.
For example: !!"he shot me down" === true and !!0 === false.

It's not a single operator; it's two. It's equivalent to the following and is a quick way to cast a value to Boolean.
val.enabled = !(!enable);

It forces all things to Boolean.
For example:
console.log(undefined); // -> undefined
console.log(!undefined); // -> true
console.log(!!undefined); // -> false
console.log('abc'); // -> abc
console.log(!'abc'); // -> false
console.log(!!'abc'); // -> true
console.log(0 === false); // -> false
console.log(!0 === false); // -> false
console.log(!!0 === false); // -> true

This question has been answered quite thoroughly, but I'd like to add an answer that I hope is as simplified as possible, making the meaning of !! as simple to grasp as can be.
Because JavaScript has what are called "truthy" and "falsy" values, there are expressions that when evaluated in other expressions will result in a true or false condition, even though the value or expression being examined is not actually true or false.
For instance:
if (document.getElementById('myElement')) {
// Code block
}
If that element does in fact exist, the expression will evaluate as true, and the code block will be executed.
However:
if (document.getElementById('myElement') == true) {
// Code block
}
...will not result in a true condition, and the code block will not be executed, even if the element does exist.
Why? Because document.getElementById() is a "truthy" expression that will evaluate as true in this if() statement, but it is not an actual Boolean value of true.
The double "not" in this case is quite simple. It is simply two nots back to back.
The first one simply "inverts" the truthy or falsy value, resulting in an actual Boolean type, and then the second one "inverts" it back again to its original state, but now in an actual Boolean value. That way you have consistency:
if (!!document.getElementById('myElement')) {}
and
if (!!document.getElementById('myElement') == true) {}
will both return true, as expected.

I suspect this is a leftover from C++ where people override the ! operator, but not the bool operator.
So to get a negative (or positive) answer in that case, you would first need to use the ! operator to get a Boolean, but if you wanted to check the positive case you would use !!.

The if and while statements and the ? operator use truth values to determine which branch of code to run. For example, zero and NaN numbers and the empty string are false, but other numbers and strings are true. Objects are true, but the undefined value and null are both false.
The double negation operator !! calculates the truth value of a value. It's actually two operators, where !!x means !(!x), and behaves as follows:
If x is a false value, !x is true, and !!x is false.
If x is a true value, !x is false, and !!x is true.
When used at the top level of a Boolean context (if, while, or ?), the !! operator is behaviorally a no-op. For example, if (x) and if (!!x) mean the same thing.
Practical uses
However it has several practical uses.
One use is to lossily compress an object to its truth value, so that your code isn't holding a reference to a big object and keeping it alive. Assigning !!some_big_object to a variable instead of some_big_object lets go of it for the garbage collector. This is useful for cases that produce either an object or a false value such as null or the undefined value, such as browser feature detection.
Another use, which I mentioned in an answer about C's corresponding !! operator, is with "lint" tools that look for common typos and print diagnostics. For example, in both C and JavaScript, a few common typos for Boolean operations produce other behaviors whose output isn't quite as Boolean:
if (a = b) is assignment followed by use of the truth value of b; if (a == b) is an equality comparison.
if (a & b) is a bitwise AND; if (a && b) is a logical AND. 2 & 5 is 0 (a false value); 2 && 5 is true.
The !! operator reassures the lint tool that what you wrote is what you meant: do this operation, then take the truth value of the result.
A third use is to produce logical XOR and logical XNOR. In both C and JavaScript, a && b performs a logical AND (true if both sides are true), and a & b performs a bitwise AND. a || b performs a logical OR (true if at least one are true), and a | b performs a bitwise OR. There's a bitwise XOR (exclusive OR) as a ^ b, but there's no built-in operator for logical XOR (true if exactly one side is true). You might, for example, want to allow the user to enter text in exactly one of two fields. What you can do is convert each to a truth value and compare them: !!x !== !!y.

I just wanted to add that
if(variableThing){
// do something
}
is the same as
if(!!variableThing){
// do something
}
But this can be an issue when something is undefined.
// a === undefined, b is an empty object (eg. b.asdf === undefined)
var a, b = {};
// Both of these give error a.foo is not defined etc.
// you'd see the same behavior for !!a.foo and !!b.foo.bar
a.foo
b.foo.bar
// This works -- these return undefined
a && a.foo
b.foo && b.foo.bar
b && b.foo && b.foo.bar
The trick here is the chain of &&s will return the first falsey value it finds -- and this can be fed to an if statement etc. So if b.foo is undefined, it will return undefined and skip the b.foo.bar statement, and we get no error.
The above return undefined but if you have an empty string, false, null, 0, undefined those values will return and soon as we encounter them in the chain -- [] and {} are both "truthy" and we will continue down the so-called "&& chain" to the next value to the right.
P.S. Another way of doing the above (b && b.foo) is (b || {}).foo. Those are equivalent, because if b is undefined then b || {} will be {}, and you'll be accessing a value in an empty object (no error) instead of trying to access a value within "undefined" (causes an error).
So, (b || {}).foo is the same as b && b.foo and ((b || {}).foo || {}).bar is the same as b && b.foo && b.foo.bar.

It is double Boolean negation. It is often used to check if a value is not undefined.

!!x is shorthand for Boolean(x).
The first bang forces the JavaScript engine to run Boolean(x), but it also has the side effect of inverting the value. So the second bang undoes the side effect.

There are tons of great answers here, but if you've read down this far, this helped me to 'get it'. Open the console in Chrome (etc.), and start typing:
!(!(1))
!(!(0))
!(!('truthy'))
!(!(null))
!(!(''))
!(!(undefined))
!(!(new Object())
!(!({}))
woo = 'hoo'
!(!(woo))
...etc., etc., until the light goes on ;)
Naturally, these are all the same as merely typing !!someThing, but the added parentheses might help make it more understandable.

After seeing all these great answers, I would like to add another reason for using !!. Currently I'm working in Angular 2-4 (TypeScript) and I want to return a Boolean as false when my user is not authenticated. If he isn't authenticated, the token-string would be null or "". I can do this by using the next block of code:
public isAuthenticated(): boolean {
return !!this.getToken();
}

It returns the Boolean value of a variable.
Instead, the Boolean class can be used.
(Please read the code descriptions.)
var X = "test"; // The X value is "test" as a String value
var booleanX = !!X // booleanX is `true` as a Boolean value because non-empty strings evaluates as `true` in Boolean
var whatIsXValueInBoolean = Boolean(X) // whatIsXValueInBoolean is `true` again
console.log(Boolean(X) === !!X) // Writes `true`
Namely, Boolean(X) = !!X in use.
Please check code snippet out below ↓
let a = 0
console.log("a: ", a) // Writes a value in its kind
console.log("!a: ", !a) // Writes '0 is NOT true in Boolean' value as Boolean - so that's true. In Boolean, 0 means false and 1 means true.
console.log("!!a: ", !!a) // Writes 0 value in Boolean. 0 means false.
console.log("Boolean(a): ", Boolean(a)) // Equals `!!a`
console.log("\n") // Newline
a = 1
console.log("a: ", a)
console.log("!a: ", !a)
console.log("!!a: ", !!a) // Writes 1 value in Boolean
console.log("\n") // Newline
a = ""
console.log("a: ", a)
console.log("!a: ", !a) // Writes '"" is NOT true in Boolean' value as Boolean - so that's true. In Boolean, empty strings, null and undefined values mean false and if there is a string it means true.
console.log("!!a: ", !!a) // Writes "" value in Boolean
console.log("\n") // Newline
a = "test"
console.log("a: ", a) // Writes a value in its kind
console.log("!a: ", !a)
console.log("!!a: ", !!a) // Writes "test" value in Boolean
console.log("Boolean(a) === !!a: ", Boolean(a) === !!a) // writes true

It is important to remember the evaluations to true and false in JavaScript:
Everything with a "Value" is true (namely truthy), for example:
101,
3.1415,
-11,
"Lucky Brain",
new Object()
and, of course, true
Everything without a "Value" is false (namely falsy), for example:
0,
-0,
"" (empty string),
undefined,
null,
NaN (not a number)
and, of course, false
Applying the "logical not" operator (!) evaluates the operand, converting it to boolean and then negating it. Applying it twice will negate the negation, effectively converting the value to boolean. Not applying the operator will just be a regular assignment of the exact value. Examples:
var value = 23; // number
var valueAsNegatedBoolean = !value; // boolean falsy (because 23 is truthy)
var valueAsBoolean = !!value; // boolean truthy
var copyOfValue = value; // number 23
var value2 = 0;
var value2AsNegatedBoolean = !value2; // boolean truthy (because 0 is falsy)
var value2AsBoolean = !!value2; // boolean falsy
var copyOfValue2 = value2; // number 0
value2 = value; assigns the exact object value even if it is not boolean hence value2 won't necessarily end up being boolean.
value2 = !!value; assigns a guaranteed boolean as the result of the double negation of the operand value and it is equivalent to the following but much shorter and readable:
if (value) {
value2 = true;
} else {
value2 = false;
}

Here is a piece of code from AngularJS:
var requestAnimationFrame = $window.requestAnimationFrame ||
$window.webkitRequestAnimationFrame ||
$window.mozRequestAnimationFrame;
var rafSupported = !!requestAnimationFrame;
Their intention is to set rafSupported to true or false based on the availability of function in requestAnimationFrame.
It can be achieved by checking in the following way in general:
if(typeof requestAnimationFrame === 'function')
rafSupported =true;
else
rafSupported =false;
The short way could be using !!
rafSupported = !!requestAnimationFrame;
So if requestAnimationFrame was assigned a function then !requestAnimationFrame would be false and one more ! of it would be true.
If requestAnimationFrame was assigned undefined then !requestAnimationFrame would be true and one more ! of it would be false.

Use the logical not operator two times.
It means !true = false and !!true = true.

Regarding null & undefined with logical js operators [duplicate]

What are the values in JavaScript that are 'falsey', meaning that they evaluate as false in expressions like if(value), value ? and !value?
There are some discussions of the purpose of falsey values on Stack Overflow already, but no exhaustive complete answer listing what all the falsey values are.
I couldn't find any complete list on MDN JavaScript Reference, and I was surprised to find that the top results when looking for a complete, authoritative list of falsey values in JavaScript were blog articles, some of which had obvious omissions (for example, NaN), and none of which had a format like Stack Overflow's where comments or alternative answers could be added to point out quirks, surprises, omissions, mistakes or caveats. So, it seemed to make sense to make one.

Falsey values in JavaScript
false
Zero of Number type: 0 and also -0, 0.0, and hex form 0x0 (thanks RBT)
Zero of BigInt type: 0n and 0x0n (new in 2020, thanks GetMeARemoteJob)
"", '' and `` - strings of length 0
null
undefined
NaN
document.all (in HTML browsers only)
This is a weird one. document.all is a falsey object, with typeof as undefined. It was a Microsoft-proprietory function in IE before IE11, and was added to the HTML spec as a "willful violation of the JavaScript specification" so that sites written for IE wouldn't break on trying to access, for example, document.all.something; it's falsy because if (document.all) used to be a popular way to detect IE, before conditional comments. See Why is document.all falsy? for details
"Falsey" simply means that JavaScript's internal ToBoolean function returns false. ToBoolean underlies !value, value ? ... : ...; and if (value). Here's its official specification (2020 working draft) (the only changes since the very first ECMAscript specification in 1997 are the addition of ES6's Symbols, which are always truthy, and BigInt, mentioned above:
Argument type
Result
Undefined
Return false.
Null
Return false.
Boolean
Return argument.
Number
If argument is +0, -0, or NaN, return false; otherwise return true.
String
If argument is the empty String (its length is zero), return false; otherwise return true.
BigInt
If argument is 0n, return false; otherwise return true.
Symbol
Return true.
Object
Return true.
Comparisons with == (loose equality)
It's worth talking about falsy values' loose comparisons with ==, which uses ToNumber() and can cause some confusion due to the underlying differences. They effectively form three groups:
false, 0, -0, "", '' all match each other with ==
e.g. false == "", '' == 0 and therefore 4/2 - 2 == 'some string'.slice(11);
null, undefined match with ==
e.g. null == undefined but undefined != false
It's also worth mentioning that while typeof null returns 'object', null is not an object, this is a longstanding bug/quirk that was not fixed in order to maintain compatibility. It's not a true object, and objects are truthy (except for that "wilful violation" document.all when Javascript is implemented in HTML)
NaN doesn't match anything, with == or ===, not even itself
e.g. NaN != NaN, NaN !== NaN, NaN != false, NaN != null
With "strict equality" (===), there are no such groupings. Only false === false.
This is one of the reasons why many developers and many style guides (e.g. standardjs) prefer === and almost never use ==.
Truthy values that actually == false
"Truthy" simply means that JavaScript's internal ToBoolean function returns true. A quirk of Javascript to be aware of (and another good reason to prefer === over ==): it is possible for a value to be truthy (ToBoolean returns true), but also == false.
You might think if (value && value == false) alert('Huh?') is a logical impossibility that couldn't happen, but it will, for:
"0" and '0' - they're non-empty strings, which are truthy, but Javascript's == matches numbers with equivalent strings (e.g. 42 == "42"). Since 0 == false, if "0" == 0, "0" == false.
new Number(0) and new Boolean(false) - they're objects, which are truthy, but == sees their values, which == false.
0 .toExponential(); - an object with a numerical value equivalent to 0
Any similar constructions that give you a false-equaling value wrapped in a type that is truthy
[], [[]] and [0] (thanks cloudfeet for the JavaScript Equality Table link)
Some more truthy values
These are just a few values that some people might expect to be falsey, but are actually truthy.
-1 and all non-zero negative numbers
' ', " ", "false", 'null'... all non-empty strings, including strings that are just whitespace
Anything from typeof, which always returns a non-empty string, for example:
typeof null (returns a string 'object' due to a longstanding bug/quirk)
typeof undefined (returns a string 'undefined')
Any object (except that "wilful violation" document.all in browsers). Remember that null isn't really an object, despite typeof suggesting otherwise. Examples:
{}
[]
function(){} or () => {} (any function, including empty functions)
Error and any instance of Error
Any regular expression
Anything created with new (including new Number(0) and new Boolean(false))
Any Symbol
true, 1, "1" and [1] return true when compared to each other with ==.

Don't forget about the non-empty string "false" which evaluates to true

Just to add to #user568458's list of falsy values:
In addition to integer number 0, the decimal number 0.0, 0.00 or any such zeroish number is also a falsy value.
var myNum = 0.0;
if(myNum){
console.log('I am a truthy value');
}
else {
console.log('I am a falsy value');
}
Above code snippet prints I am a falsy value
Similarly hex representation of the number 0 is also a falsy value as shown in below code snippet:
var myNum = 0x0; //hex representation of 0
if(myNum){
console.log('I am a truthy value');
}
else {
console.log('I am a falsy value');
}
Above code snippet again prints I am a falsy value.

Addition to the topic, as of ES2020 we have a new value which is falsy, it's BigInt zero (0n):
0n == false // true
-0n == false // true
0n === false // false
-0n === false // false
So with this, we now have 7 "falsy" values in total (not including document.all as mentioned by user above since it's part of DOM and not JS).

Javascript "implicit" condition in IF statements [duplicate]

What are the values in JavaScript that are 'falsey', meaning that they evaluate as false in expressions like if(value), value ? and !value?
There are some discussions of the purpose of falsey values on Stack Overflow already, but no exhaustive complete answer listing what all the falsey values are.
I couldn't find any complete list on MDN JavaScript Reference, and I was surprised to find that the top results when looking for a complete, authoritative list of falsey values in JavaScript were blog articles, some of which had obvious omissions (for example, NaN), and none of which had a format like Stack Overflow's where comments or alternative answers could be added to point out quirks, surprises, omissions, mistakes or caveats. So, it seemed to make sense to make one.

Falsey values in JavaScript
false
Zero of Number type: 0 and also -0, 0.0, and hex form 0x0 (thanks RBT)
Zero of BigInt type: 0n and 0x0n (new in 2020, thanks GetMeARemoteJob)
"", '' and `` - strings of length 0
null
undefined
NaN
document.all (in HTML browsers only)
This is a weird one. document.all is a falsey object, with typeof as undefined. It was a Microsoft-proprietory function in IE before IE11, and was added to the HTML spec as a "willful violation of the JavaScript specification" so that sites written for IE wouldn't break on trying to access, for example, document.all.something; it's falsy because if (document.all) used to be a popular way to detect IE, before conditional comments. See Why is document.all falsy? for details
"Falsey" simply means that JavaScript's internal ToBoolean function returns false. ToBoolean underlies !value, value ? ... : ...; and if (value). Here's its official specification (2020 working draft) (the only changes since the very first ECMAscript specification in 1997 are the addition of ES6's Symbols, which are always truthy, and BigInt, mentioned above:
Argument type
Result
Undefined
Return false.
Null
Return false.
Boolean
Return argument.
Number
If argument is +0, -0, or NaN, return false; otherwise return true.
String
If argument is the empty String (its length is zero), return false; otherwise return true.
BigInt
If argument is 0n, return false; otherwise return true.
Symbol
Return true.
Object
Return true.
Comparisons with == (loose equality)
It's worth talking about falsy values' loose comparisons with ==, which uses ToNumber() and can cause some confusion due to the underlying differences. They effectively form three groups:
false, 0, -0, "", '' all match each other with ==
e.g. false == "", '' == 0 and therefore 4/2 - 2 == 'some string'.slice(11);
null, undefined match with ==
e.g. null == undefined but undefined != false
It's also worth mentioning that while typeof null returns 'object', null is not an object, this is a longstanding bug/quirk that was not fixed in order to maintain compatibility. It's not a true object, and objects are truthy (except for that "wilful violation" document.all when Javascript is implemented in HTML)
NaN doesn't match anything, with == or ===, not even itself
e.g. NaN != NaN, NaN !== NaN, NaN != false, NaN != null
With "strict equality" (===), there are no such groupings. Only false === false.
This is one of the reasons why many developers and many style guides (e.g. standardjs) prefer === and almost never use ==.
Truthy values that actually == false
"Truthy" simply means that JavaScript's internal ToBoolean function returns true. A quirk of Javascript to be aware of (and another good reason to prefer === over ==): it is possible for a value to be truthy (ToBoolean returns true), but also == false.
You might think if (value && value == false) alert('Huh?') is a logical impossibility that couldn't happen, but it will, for:
"0" and '0' - they're non-empty strings, which are truthy, but Javascript's == matches numbers with equivalent strings (e.g. 42 == "42"). Since 0 == false, if "0" == 0, "0" == false.
new Number(0) and new Boolean(false) - they're objects, which are truthy, but == sees their values, which == false.
0 .toExponential(); - an object with a numerical value equivalent to 0
Any similar constructions that give you a false-equaling value wrapped in a type that is truthy
[], [[]] and [0] (thanks cloudfeet for the JavaScript Equality Table link)
Some more truthy values
These are just a few values that some people might expect to be falsey, but are actually truthy.
-1 and all non-zero negative numbers
' ', " ", "false", 'null'... all non-empty strings, including strings that are just whitespace
Anything from typeof, which always returns a non-empty string, for example:
typeof null (returns a string 'object' due to a longstanding bug/quirk)
typeof undefined (returns a string 'undefined')
Any object (except that "wilful violation" document.all in browsers). Remember that null isn't really an object, despite typeof suggesting otherwise. Examples:
{}
[]
function(){} or () => {} (any function, including empty functions)
Error and any instance of Error
Any regular expression
Anything created with new (including new Number(0) and new Boolean(false))
Any Symbol
true, 1, "1" and [1] return true when compared to each other with ==.

Don't forget about the non-empty string "false" which evaluates to true

Just to add to #user568458's list of falsy values:
In addition to integer number 0, the decimal number 0.0, 0.00 or any such zeroish number is also a falsy value.
var myNum = 0.0;
if(myNum){
console.log('I am a truthy value');
}
else {
console.log('I am a falsy value');
}
Above code snippet prints I am a falsy value
Similarly hex representation of the number 0 is also a falsy value as shown in below code snippet:
var myNum = 0x0; //hex representation of 0
if(myNum){
console.log('I am a truthy value');
}
else {
console.log('I am a falsy value');
}
Above code snippet again prints I am a falsy value.

Addition to the topic, as of ES2020 we have a new value which is falsy, it's BigInt zero (0n):
0n == false // true
-0n == false // true
0n === false // false
-0n === false // false
So with this, we now have 7 "falsy" values in total (not including document.all as mentioned by user above since it's part of DOM and not JS).

Javascript falsy values (null, undefined, false, empty string: "" or '' and 0) and comparison(==) operator [duplicate]

This question already has answers here:
JavaScript: What is the difference between `if (!x)` and `if (x == null)`?
(5 answers)
Closed 9 years ago.
When I am using any one of values(null, undefined, false, '', 0) in a if statement, it is always evaluated as fallacy(false). Also, the negation of these values((null, undefined, false, '', 0) in a if statement always evaluated as tautology(true).
if(null){
}else{
}
if(undefined){
}else{
}
if(false){
}else{
}
if(''){
}else{
}
if(0){
}else{
}
In all the above cases, if statement is evaluated as false & else statement executes.
However, when I am comparing these fallacy values with == operator, it is not returning true always. Surprisingly, it is always returning true values when I am comparing the negation of these values.
if double equalto (==) operator checks/compares for values & not strictly for types, then why:
null == false // returns false
null == 0 // returns false
null == '' // returns false
But,
!null == !false // returns true
!null == !0 // returns true
!false == !undefined // returns true
And,
null == undefined // returns true
false == 0 // returns true
I appreciate if any one can clarify the behavior or relationship among these values(null, undefined, false, '', 0).

A common misconception
"...If double equalto (==) operator only checks/compares for values & not for types..."
That's an incorrect assumption, though it's often repeated by people. In reality, the == does check types, and in fact pays far more attention to the types than a === comparison does.
See Abstract Equality Comparison Algorithm.
A == comparison doesn't do a simple toBoolean conversion. Rather it walks through a somewhat complex recursive algorithm, which, after checking the types, attempts to coerce the operands to the same type if they don't match.
The type coercion that it performs is very specific to the types of the operands. A different sequence of coercions can take place for different type pairs. Usually (but not always) it ends up ultimately coercing the operands down to number types.
Why !ing the operands changes things
When you manually coerce both operands using !, you're now doing a simple toBoolean conversion causing the types to match, which avoids the type coercive part of the algorithm, making it behave essentially like the Strict Equality Comparison Algorithm.
So the only way to predict the outcome of a == comparison when the types don't match is to understand that Abstract algorithm.
Don't forget about NaN
And FYI, there's one more "falsey" value to consider, NaN. Its == comparison will always be false, no matter what. Even when comparing to another NaN value, it'll be false.

undefined: means a variable was declared but has no value assigned
null: the value of null has been assigned, which means it has no value
false, '' and 0 I think you can probably work out what these mean.

NULL is different from false (NULL is of type object and false is of type of boolean), null is different from 0 (0 is of type integer), null is also different from '' ('' is of type of string). but they are all falsy values. The ! operator negates a boolean value. If ! is used on falsy values, it results to conversion of falsy values to an object of type boolean.

Categories