We Keep Coding

Which function is better to cover most cases to check if a data is Object or not between 2 function below

I'm usually using this condition to check if a value of data is Object or not:
// function A
export const isObject = item => {
return item && item.constructor === Object;
};
since it is easier to understand, but my friend usually using this function to check data with the same condition as above isObject(value) {}:
// function B
export function isObject(value) {
return typeof value === "object" && !Array.isArray(value) && value !== null;
}
we got confused when there is a conflict when we want to merge to latest branch, so here is the question, Between function A & B which one is better to cover most case regarding object data?, Can you explain it too so I have a better knowledge about it.

Assuming that you want to identify plain objects that inherit from Object.prototype, and aren't subclasses of something else:
The first approach could be fooled because the .constructor property doesn't necessarily refer to the constructor function used to create the function - an object could assign something else to that property.
const isObject = item => {
return item && item.constructor === Object;
};
const a = [];
console.log(a.constructor == Array);
a.constructor = Object;
console.log(a.constructor === Object);
console.log(isObject(a));
The second approach could be fooled because there are other sorts of things than arrays that inherit from Object.prototype that aren't plain objects nor arrays - like HTMLElements and many, many other things.
function isObject(value) {
return typeof value === "object" && !Array.isArray(value) && value !== null;
}
console.log(isObject(document.body));
console.log(isObject(new Proxy({}, {})));
So I prefer Object.getPrototypeOf, with which you can be quite confident that the value's immediate internal prototype is Object.prototype, and that it's nothing else special.
const isObject = item => Object.getPrototypeOf(item) === Object.prototype;
console.log(isObject([]));
console.log(isObject({}));
console.log(isObject(document));
It could fail if Object.getPrototypeOf is overwritten, of course - but if you're at that point, pretty much nothing is trustworthy anyway.
Note that many things inherit from Object.prototype which will give true for this, such as Number.prototype - if you want to exclude those kinds, you could check if a .constructor property exists.

How to check if an object exists before getting value? [duplicate]

In my code, I deal with an array that has some entries with many objects nested inside one another, where as some do not. It looks something like the following:
// where this array is hundreds of entries long, with a mix
// of the two examples given
var test = [{'a':{'b':{'c':"foo"}}}, {'a': "bar"}];
This is giving me problems because I need to iterate through the array at times, and the inconsistency is throwing me errors like so:
for (i=0; i<test.length; i++) {
// ok on i==0, but 'cannot read property of undefined' on i==1
console.log(a.b.c);
}
I am aware that I can say if(a.b){ console.log(a.b.c)}, but this is extraordinarily tedious in cases where there are up to 5 or 6 objects nested within one another. Is there any other (easier) way that I can have it ONLY do the console.log if it exists, but without throwing an error?

Update:
If you use JavaScript according to ECMAScript 2020 or later, see optional chaining.
TypeScript has added support for optional chaining in version 3.7.
// use it like this
obj?.a?.lot?.of?.properties
Solution for JavaScript before ECMASCript 2020 or TypeScript older than version 3.7:
A quick workaround is using a try/catch helper function with ES6 arrow function:
function getSafe(fn, defaultVal) {
try {
return fn();
} catch (e) {
return defaultVal;
}
}
// use it like this
console.log(getSafe(() => obj.a.lot.of.properties));
// or add an optional default value
console.log(getSafe(() => obj.a.lot.of.properties, 'nothing'));

What you are doing raises an exception (and rightfully so).
You can always do:
try{
window.a.b.c
}catch(e){
console.log("YO",e)
}
But I wouldn't, instead think of your use case.
Why are you accessing data, 6 levels nested that you are unfamiliar of? What use case justifies this?
Usually, you'd like to actually validate what sort of object you're dealing with.
Also, on a side note you should not use statements like if(a.b) because it will return false if a.b is 0 or even if it is "0". Instead check if a.b !== undefined

If I am understanding your question correctly, you want the safest way to determine if an object contains a property.
The easiest way is to use the in operator.
window.a = "aString";
//window should have 'a' property
//lets test if it exists
if ("a" in window){
//true
}
if ("b" in window){
//false
}
Of course you can nest this as deep as you want
if ("a" in window.b.c) { }
Not sure if this helps.

Try this. If a.b is undefined, it will leave the if statement without any exception.
if (a.b && a.b.c) {
console.log(a.b.c);
}

If you are using lodash, you could use their has function. It is similar to the native "in", but allows paths.
var testObject = {a: {b: {c: 'walrus'}}};
if(_.has(testObject, 'a.b.c')) {
//Safely access your walrus here
}

If you use Babel, you can already use the optional chaining syntax with #babel/plugin-proposal-optional-chaining Babel plugin. This would allow you to replace this:
console.log(a && a.b && a.b.c);
with this:
console.log(a?.b?.c);

If you have lodash you can use its .get method
_.get(a, 'b.c.d.e')
or give it a default value
_.get(a, 'b.c.d.e', default)

I use undefsafe religiously. It tests each level down into your object until it either gets the value you asked for, or it returns "undefined". But never errors.

This is a common issue when working with deep or complex json object, so I try to avoid try/catch or embedding multiple checks which would make the code unreadable, I usually use this little piece of code in all my procect to do the job.
/* ex: getProperty(myObj,'aze.xyz',0) // return myObj.aze.xyz safely
* accepts array for property names:
* getProperty(myObj,['aze','xyz'],{value: null})
*/
function getProperty(obj, props, defaultValue) {
var res, isvoid = function(x){return typeof x === "undefined" || x === null;}
if(!isvoid(obj)){
if(isvoid(props)) props = [];
if(typeof props === "string") props = props.trim().split(".");
if(props.constructor === Array){
res = props.length>1 ? getProperty(obj[props.shift()],props,defaultValue) : obj[props[0]];
}
}
return typeof res === "undefined" ? defaultValue: res;
}

I like Cao Shouguang's answer, but I am not fond of passing a function as parameter into the getSafe function each time I do the call. I have modified the getSafe function to accept simple parameters and pure ES5.
/**
* Safely get object properties.
* #param {*} prop The property of the object to retrieve
* #param {*} defaultVal The value returned if the property value does not exist
* #returns If property of object exists it is returned,
* else the default value is returned.
* #example
* var myObj = {a : {b : 'c'} };
* var value;
*
* value = getSafe(myObj.a.b,'No Value'); //returns c
* value = getSafe(myObj.a.x,'No Value'); //returns 'No Value'
*
* if (getSafe(myObj.a.x, false)){
* console.log('Found')
* } else {
* console.log('Not Found')
* }; //logs 'Not Found'
*
* if(value = getSafe(myObj.a.b, false)){
* console.log('New Value is', value); //logs 'New Value is c'
* }
*/
function getSafe(prop, defaultVal) {
return function(fn, defaultVal) {
try {
if (fn() === undefined) {
return defaultVal;
} else {
return fn();
}
} catch (e) {
return defaultVal;
}
}(function() {return prop}, defaultVal);
}

Lodash has a get method which allows for a default as an optional third parameter, as show below:
const myObject = {
has: 'some',
missing: {
vars: true
}
}
const path = 'missing.const.value';
const myValue = _.get(myObject, path, 'default');
console.log(myValue) // prints out default, which is specified above
<script src="https://cdnjs.cloudflare.com/ajax/libs/lodash.js/4.17.11/lodash.js"></script>

Imagine that we want to apply a series of functions to x if and only if x is non-null:
if (x !== null) x = a(x);
if (x !== null) x = b(x);
if (x !== null) x = c(x);
Now let's say that we need to do the same to y:
if (y !== null) y = a(y);
if (y !== null) y = b(y);
if (y !== null) y = c(y);
And the same to z:
if (z !== null) z = a(z);
if (z !== null) z = b(z);
if (z !== null) z = c(z);
As you can see without a proper abstraction, we'll end up duplicating code over and over again. Such an abstraction already exists: the Maybe monad.
The Maybe monad holds both a value and a computational context:
The monad keeps the value safe and applies functions to it.
The computational context is a null check before applying a function.
A naive implementation would look like this:
⚠️ This implementation is for illustration purpose only! This is not how it should be done and is wrong at many levels. However this should give you a better idea of what I am talking about.
As you can see nothing can break:
We apply a series of functions to our value
If at any point, the value becomes null (or undefined) we just don't apply any function anymore.
const abc = obj =>
Maybe
.of(obj)
.map(o => o.a)
.map(o => o.b)
.map(o => o.c)
.value;
const values = [
{},
{a: {}},
{a: {b: {}}},
{a: {b: {c: 42}}}
];
console.log(
values.map(abc)
);
<script>
function Maybe(x) {
this.value = x; //-> container for our value
}
Maybe.of = x => new Maybe(x);
Maybe.prototype.map = function (fn) {
if (this.value == null) { //-> computational context
return this;
}
return Maybe.of(fn(this.value));
};
</script>
Appendix 1
I cannot explain what monads are as this is not the purpose of this post and there are people out there better at this than I am. However as Eric Elliot said in hist blog post JavaScript Monads Made Simple:
Regardless of your skill level or understanding of category theory, using monads makes your code easier to work with. Failing to take advantage of monads may make your code harder to work with (e.g., callback hell, nested conditional branches, more verbosity).
Appendix 2
Here's how I'd solve your issue using the Maybe monad from monetjs
const prop = key => obj => Maybe.fromNull(obj[key]);
const abc = obj =>
Maybe
.fromNull(obj)
.flatMap(prop('a'))
.flatMap(prop('b'))
.flatMap(prop('c'))
.orSome('🌯')
const values = [
{},
{a: {}},
{a: {b: {}}},
{a: {b: {c: 42}}}
];
console.log(
values.map(abc)
);
<script src="https://www.unpkg.com/monet#0.9.0/dist/monet.js"></script>
<script>const {Maybe} = Monet;</script>

In str's answer, value 'undefined' will be returned instead of the set default value if the property is undefined. This sometimes can cause bugs. The following will make sure the defaultVal will always be returned when either the property or the object is undefined.
const temp = {};
console.log(getSafe(()=>temp.prop, '0'));
function getSafe(fn, defaultVal) {
try {
if (fn() === undefined || fn() === null) {
return defaultVal
} else {
return fn();
}
} catch (e) {
return defaultVal;
}
}

You can use optional chaining from the ECMAScript standart.
Like this:
a?.b?.c?.d?.func?.()

I answered this before and happened to be doing a similar check today. A simplification to check if a nested dotted property exists. You could modify this to return the value, or some default to accomplish your goal.
function containsProperty(instance, propertyName) {
// make an array of properties to walk through because propertyName can be nested
// ex "test.test2.test.test"
let walkArr = propertyName.indexOf('.') > 0 ? propertyName.split('.') : [propertyName];
// walk the tree - if any property does not exist then return false
for (let treeDepth = 0, maxDepth = walkArr.length; treeDepth < maxDepth; treeDepth++) {
// property does not exist
if (!Object.prototype.hasOwnProperty.call(instance, walkArr[treeDepth])) {
return false;
}
// does it exist - reassign the leaf
instance = instance[walkArr[treeDepth]];
}
// default
return true;
}
In your question you could do something like:
let test = [{'a':{'b':{'c':"foo"}}}, {'a': "bar"}];
containsProperty(test[0], 'a.b.c');

I usually use like this:
var x = object.any ? object.any.a : 'def';

You can avoid getting an error by giving a default value before getting the property
var test = [{'a':{'b':{'c':"foo"}}}, {'a': "bar"}];
for (i=0; i<test.length; i++) {
const obj = test[i]
// No error, just undefined, which is ok
console.log(((obj.a || {}).b || {}).c);
}
This works great with arrays too:
const entries = [{id: 1, name: 'Scarllet'}]
// Giving a default name when is empty
const name = (entries.find(v => v.id === 100) || []).name || 'no-name'
console.log(name)

Unrelated to the question's actual question, but might be useful for people coming to this question looking for answers.
Check your function parameters.
If you have a function like const x({ a }) => { }, and you call it without arguments x(); append = {} to the parameter: const x({ a } = {}) => { }.
What I had
I had a function like this:
const x = ({ a }) => console.log(a);
// This one works as expected
x({ a: 1 });
// This one errors out
x();
Which results in "Uncaught TypeError: Cannot destructure property 'a' of 'undefined' as it is undefined."
What I switched it to (now works).
const x = ({ a } = {}) => console.log(a);
// This one works as expected
x({ a: 1 });
// This now works too!
x();

How to pass props from a container component to a class component [duplicate]

In my code, I deal with an array that has some entries with many objects nested inside one another, where as some do not. It looks something like the following:
// where this array is hundreds of entries long, with a mix
// of the two examples given
var test = [{'a':{'b':{'c':"foo"}}}, {'a': "bar"}];
This is giving me problems because I need to iterate through the array at times, and the inconsistency is throwing me errors like so:
for (i=0; i<test.length; i++) {
// ok on i==0, but 'cannot read property of undefined' on i==1
console.log(a.b.c);
}
I am aware that I can say if(a.b){ console.log(a.b.c)}, but this is extraordinarily tedious in cases where there are up to 5 or 6 objects nested within one another. Is there any other (easier) way that I can have it ONLY do the console.log if it exists, but without throwing an error?

Update:
If you use JavaScript according to ECMAScript 2020 or later, see optional chaining.
TypeScript has added support for optional chaining in version 3.7.
// use it like this
obj?.a?.lot?.of?.properties
Solution for JavaScript before ECMASCript 2020 or TypeScript older than version 3.7:
A quick workaround is using a try/catch helper function with ES6 arrow function:
function getSafe(fn, defaultVal) {
try {
return fn();
} catch (e) {
return defaultVal;
}
}
// use it like this
console.log(getSafe(() => obj.a.lot.of.properties));
// or add an optional default value
console.log(getSafe(() => obj.a.lot.of.properties, 'nothing'));

What you are doing raises an exception (and rightfully so).
You can always do:
try{
window.a.b.c
}catch(e){
console.log("YO",e)
}
But I wouldn't, instead think of your use case.
Why are you accessing data, 6 levels nested that you are unfamiliar of? What use case justifies this?
Usually, you'd like to actually validate what sort of object you're dealing with.
Also, on a side note you should not use statements like if(a.b) because it will return false if a.b is 0 or even if it is "0". Instead check if a.b !== undefined

If I am understanding your question correctly, you want the safest way to determine if an object contains a property.
The easiest way is to use the in operator.
window.a = "aString";
//window should have 'a' property
//lets test if it exists
if ("a" in window){
//true
}
if ("b" in window){
//false
}
Of course you can nest this as deep as you want
if ("a" in window.b.c) { }
Not sure if this helps.

Try this. If a.b is undefined, it will leave the if statement without any exception.
if (a.b && a.b.c) {
console.log(a.b.c);
}

If you are using lodash, you could use their has function. It is similar to the native "in", but allows paths.
var testObject = {a: {b: {c: 'walrus'}}};
if(_.has(testObject, 'a.b.c')) {
//Safely access your walrus here
}

If you use Babel, you can already use the optional chaining syntax with #babel/plugin-proposal-optional-chaining Babel plugin. This would allow you to replace this:
console.log(a && a.b && a.b.c);
with this:
console.log(a?.b?.c);

If you have lodash you can use its .get method
_.get(a, 'b.c.d.e')
or give it a default value
_.get(a, 'b.c.d.e', default)

I use undefsafe religiously. It tests each level down into your object until it either gets the value you asked for, or it returns "undefined". But never errors.

This is a common issue when working with deep or complex json object, so I try to avoid try/catch or embedding multiple checks which would make the code unreadable, I usually use this little piece of code in all my procect to do the job.
/* ex: getProperty(myObj,'aze.xyz',0) // return myObj.aze.xyz safely
* accepts array for property names:
* getProperty(myObj,['aze','xyz'],{value: null})
*/
function getProperty(obj, props, defaultValue) {
var res, isvoid = function(x){return typeof x === "undefined" || x === null;}
if(!isvoid(obj)){
if(isvoid(props)) props = [];
if(typeof props === "string") props = props.trim().split(".");
if(props.constructor === Array){
res = props.length>1 ? getProperty(obj[props.shift()],props,defaultValue) : obj[props[0]];
}
}
return typeof res === "undefined" ? defaultValue: res;
}

I like Cao Shouguang's answer, but I am not fond of passing a function as parameter into the getSafe function each time I do the call. I have modified the getSafe function to accept simple parameters and pure ES5.
/**
* Safely get object properties.
* #param {*} prop The property of the object to retrieve
* #param {*} defaultVal The value returned if the property value does not exist
* #returns If property of object exists it is returned,
* else the default value is returned.
* #example
* var myObj = {a : {b : 'c'} };
* var value;
*
* value = getSafe(myObj.a.b,'No Value'); //returns c
* value = getSafe(myObj.a.x,'No Value'); //returns 'No Value'
*
* if (getSafe(myObj.a.x, false)){
* console.log('Found')
* } else {
* console.log('Not Found')
* }; //logs 'Not Found'
*
* if(value = getSafe(myObj.a.b, false)){
* console.log('New Value is', value); //logs 'New Value is c'
* }
*/
function getSafe(prop, defaultVal) {
return function(fn, defaultVal) {
try {
if (fn() === undefined) {
return defaultVal;
} else {
return fn();
}
} catch (e) {
return defaultVal;
}
}(function() {return prop}, defaultVal);
}

Lodash has a get method which allows for a default as an optional third parameter, as show below:
const myObject = {
has: 'some',
missing: {
vars: true
}
}
const path = 'missing.const.value';
const myValue = _.get(myObject, path, 'default');
console.log(myValue) // prints out default, which is specified above
<script src="https://cdnjs.cloudflare.com/ajax/libs/lodash.js/4.17.11/lodash.js"></script>

Imagine that we want to apply a series of functions to x if and only if x is non-null:
if (x !== null) x = a(x);
if (x !== null) x = b(x);
if (x !== null) x = c(x);
Now let's say that we need to do the same to y:
if (y !== null) y = a(y);
if (y !== null) y = b(y);
if (y !== null) y = c(y);
And the same to z:
if (z !== null) z = a(z);
if (z !== null) z = b(z);
if (z !== null) z = c(z);
As you can see without a proper abstraction, we'll end up duplicating code over and over again. Such an abstraction already exists: the Maybe monad.
The Maybe monad holds both a value and a computational context:
The monad keeps the value safe and applies functions to it.
The computational context is a null check before applying a function.
A naive implementation would look like this:
⚠️ This implementation is for illustration purpose only! This is not how it should be done and is wrong at many levels. However this should give you a better idea of what I am talking about.
As you can see nothing can break:
We apply a series of functions to our value
If at any point, the value becomes null (or undefined) we just don't apply any function anymore.
const abc = obj =>
Maybe
.of(obj)
.map(o => o.a)
.map(o => o.b)
.map(o => o.c)
.value;
const values = [
{},
{a: {}},
{a: {b: {}}},
{a: {b: {c: 42}}}
];
console.log(
values.map(abc)
);
<script>
function Maybe(x) {
this.value = x; //-> container for our value
}
Maybe.of = x => new Maybe(x);
Maybe.prototype.map = function (fn) {
if (this.value == null) { //-> computational context
return this;
}
return Maybe.of(fn(this.value));
};
</script>
Appendix 1
I cannot explain what monads are as this is not the purpose of this post and there are people out there better at this than I am. However as Eric Elliot said in hist blog post JavaScript Monads Made Simple:
Regardless of your skill level or understanding of category theory, using monads makes your code easier to work with. Failing to take advantage of monads may make your code harder to work with (e.g., callback hell, nested conditional branches, more verbosity).
Appendix 2
Here's how I'd solve your issue using the Maybe monad from monetjs
const prop = key => obj => Maybe.fromNull(obj[key]);
const abc = obj =>
Maybe
.fromNull(obj)
.flatMap(prop('a'))
.flatMap(prop('b'))
.flatMap(prop('c'))
.orSome('🌯')
const values = [
{},
{a: {}},
{a: {b: {}}},
{a: {b: {c: 42}}}
];
console.log(
values.map(abc)
);
<script src="https://www.unpkg.com/monet#0.9.0/dist/monet.js"></script>
<script>const {Maybe} = Monet;</script>

In str's answer, value 'undefined' will be returned instead of the set default value if the property is undefined. This sometimes can cause bugs. The following will make sure the defaultVal will always be returned when either the property or the object is undefined.
const temp = {};
console.log(getSafe(()=>temp.prop, '0'));
function getSafe(fn, defaultVal) {
try {
if (fn() === undefined || fn() === null) {
return defaultVal
} else {
return fn();
}
} catch (e) {
return defaultVal;
}
}

You can use optional chaining from the ECMAScript standart.
Like this:
a?.b?.c?.d?.func?.()

I answered this before and happened to be doing a similar check today. A simplification to check if a nested dotted property exists. You could modify this to return the value, or some default to accomplish your goal.
function containsProperty(instance, propertyName) {
// make an array of properties to walk through because propertyName can be nested
// ex "test.test2.test.test"
let walkArr = propertyName.indexOf('.') > 0 ? propertyName.split('.') : [propertyName];
// walk the tree - if any property does not exist then return false
for (let treeDepth = 0, maxDepth = walkArr.length; treeDepth < maxDepth; treeDepth++) {
// property does not exist
if (!Object.prototype.hasOwnProperty.call(instance, walkArr[treeDepth])) {
return false;
}
// does it exist - reassign the leaf
instance = instance[walkArr[treeDepth]];
}
// default
return true;
}
In your question you could do something like:
let test = [{'a':{'b':{'c':"foo"}}}, {'a': "bar"}];
containsProperty(test[0], 'a.b.c');

I usually use like this:
var x = object.any ? object.any.a : 'def';

You can avoid getting an error by giving a default value before getting the property
var test = [{'a':{'b':{'c':"foo"}}}, {'a': "bar"}];
for (i=0; i<test.length; i++) {
const obj = test[i]
// No error, just undefined, which is ok
console.log(((obj.a || {}).b || {}).c);
}
This works great with arrays too:
const entries = [{id: 1, name: 'Scarllet'}]
// Giving a default name when is empty
const name = (entries.find(v => v.id === 100) || []).name || 'no-name'
console.log(name)

Unrelated to the question's actual question, but might be useful for people coming to this question looking for answers.
Check your function parameters.
If you have a function like const x({ a }) => { }, and you call it without arguments x(); append = {} to the parameter: const x({ a } = {}) => { }.
What I had
I had a function like this:
const x = ({ a }) => console.log(a);
// This one works as expected
x({ a: 1 });
// This one errors out
x();
Which results in "Uncaught TypeError: Cannot destructure property 'a' of 'undefined' as it is undefined."
What I switched it to (now works).
const x = ({ a } = {}) => console.log(a);
// This one works as expected
x({ a: 1 });
// This now works too!
x();

Initializing an array in Javascript returns an object [duplicate]

This question already has answers here:
How do I check if a variable is an array in JavaScript?
(24 answers)
Closed 1 year ago.
I'm trying to write a function that either accepts a list of strings, or a single string. If it's a string, then I want to convert it to an array with just the one item so I can loop over it without fear of an error.
So how do I check if the variable is an array?

The method given in the ECMAScript standard to find the class of Object is to use the toString method from Object.prototype.
if(Object.prototype.toString.call(someVar) === '[object Array]') {
alert('Array!');
}
Or you could use typeof to test if it is a string:
if(typeof someVar === 'string') {
someVar = [someVar];
}
Or if you're not concerned about performance, you could just do a concat to a new empty Array.
someVar = [].concat(someVar);
There's also the constructor which you can query directly:
if (somevar.constructor.name == "Array") {
// do something
}
Check out a thorough treatment from T.J. Crowder's blog, as posted in his comment below.
Check out this benchmark to get an idea which method performs better: http://jsben.ch/#/QgYAV
From #Bharath, convert a string to an array using ES6 for the question asked:
const convertStringToArray = (object) => {
return (typeof object === 'string') ? Array(object) : object
}
Suppose:
let m = 'bla'
let n = ['bla','Meow']
let y = convertStringToArray(m)
let z = convertStringToArray(n)
console.log('check y: '+JSON.stringify(y)) . // check y: ['bla']
console.log('check y: '+JSON.stringify(z)) . // check y: ['bla','Meow']

In modern browsers you can do:
Array.isArray(obj)
(Supported by Chrome 5, Firefox 4.0, Internet Explorer 9, Opera 10.5 and Safari 5)
For backward compatibility you can add the following:
// Only implement if no native implementation is available
if (typeof Array.isArray === 'undefined') {
Array.isArray = function(obj) {
return Object.prototype.toString.call(obj) === '[object Array]';
}
};
If you use jQuery you can use jQuery.isArray(obj) or $.isArray(obj). If you use Underscore.js you can use _.isArray(obj).
If you don't need to detect arrays created in different frames you can also just use instanceof:
obj instanceof Array

I would first check if your implementation supports isArray:
if (Array.isArray)
return Array.isArray(v);
You could also try using the instanceof operator
v instanceof Array

jQuery also offers an $.isArray() method:
var a = ["A", "AA", "AAA"];
if($.isArray(a)) {
alert("a is an array!");
} else {
alert("a is not an array!");
}
<script src="https://ajax.googleapis.com/ajax/libs/jquery/2.1.1/jquery.min.js"></script>

This is the fastest among all methods (all browsers supported):
function isArray(obj){
return !!obj && obj.constructor === Array;
}

Imagine you have this array below:
var arr = [1,2,3,4,5];
JavaScript (new and older browsers):
function isArray(arr) {
return arr.constructor.toString().indexOf("Array") > -1;
}
or
function isArray(arr) {
return arr instanceof Array;
}
or
function isArray(arr) {
return Object.prototype.toString.call(arr) === '[object Array]';
}
Then call it like this:
isArray(arr);
JavaScript (Internet Explorer 9+, Chrome 5+, Firefox 4+, Safari 5+, and Opera 10.5+)
Array.isArray(arr);
jQuery:
$.isArray(arr);
Angular:
angular.isArray(arr);
Underscore.js and Lodash:
_.isArray(arr);

Array.isArray works fast, but it isn't supported by all versions of browsers.
So you could make an exception for others and use a universal method:
Utils = {};
Utils.isArray = ('isArray' in Array) ?
Array.isArray :
function (value) {
return Object.prototype.toString.call(value) === '[object Array]';
}

A simple function to check this:
function isArray(object)
{
return object.constructor === Array;
}

You can use Array.isArray(). Here is a polyfill:
if (Array.isArray == null) {
Array.isArray = (arr) => Object.prototype.toString.call(arr) === "[object Array]"
}

As MDN says in here:
use Array.isArray or Object.prototype.toString.call to differentiate
regular objects from arrays
Like this:
Object.prototype.toString.call(arr) === '[object Array]', or
Array.isArray(arr)

There's just one line solution for this question
x instanceof Array
where x is the variable it will return true if x is an array and false if it is not.

I would make a function to test the type of object you are dealing with...
function whatAmI(me){ return Object.prototype.toString.call(me).split(/\W/)[2]; }
// tests
console.log(
whatAmI(["aiming","#"]),
whatAmI({living:4,breathing:4}),
whatAmI(function(ing){ return ing+" to the global window" }),
whatAmI("going to do with you?")
);
// output: Array Object Function String
then you can write a simple if statement...
if(whatAmI(myVar) === "Array"){
// do array stuff
} else { // could also check `if(whatAmI(myVar) === "String")` here to be sure
// do string stuff
}

You can check the type of your variable whether it is an array with;
var myArray=[];
if(myArray instanceof Array)
{
....
}

I do this in a very simple way. It works for me.
Array.prototype.isArray = true;
a=[]; b={};
a.isArray // true
b.isArray // (undefined -> false)

This is my attempt to improve on this answer taking into account the comments:
var isArray = myArray && myArray.constructor === Array;
It gets rid of the if/else, and accounts for the possibility of the array being null or undefined

I have updated the jsperf fiddle with two alternative methods as well as error checking.
It turns out that the method defining a constant value in the 'Object' and 'Array' prototypes is faster than any of the other methods. It is a somewhat surprising result.
/* Initialisation */
Object.prototype.isArray = function() {
return false;
};
Array.prototype.isArray = function() {
return true;
};
Object.prototype._isArray = false;
Array.prototype._isArray = true;
var arr = ["1", "2"];
var noarr = "1";
/* Method 1 (function) */
if (arr.isArray()) document.write("arr is an array according to function<br/>");
if (!noarr.isArray()) document.write("noarr is not an array according to function<br/>");
/* Method 2 (value) - **** FASTEST ***** */
if (arr._isArray) document.write("arr is an array according to member value<br/>");
if (!noarr._isArray) document.write("noarr is not an array according to member value<br/>");
These two methods do not work if the variable takes the undefined value, but they do work if you are certain that they have a value. With regards to checking with performance in mind if a value is an array or a single value, the second method looks like a valid fast method. It is slightly faster than 'instanceof' on Chrome, twice as fast as the second best method in Internet Explorer, Opera and Safari (on my machine).

I know, that people are looking for some kind of raw JavaScript approach. But if you want think less about it, take a look at Underscore.js' isArray:
_.isArray(object)
It returns true if object is an Array.
(function(){ return _.isArray(arguments); })();
=> false
_.isArray([1,2,3]);
=> true

The best practice is to compare it using constructor, something like this
if(some_variable.constructor === Array){
// do something
}
You can use other methods too, like typeOf, converting it to a string and then comparing, but comparing it with dataType is always a better approach.

The best solution I've seen is a cross-browser replacement for typeof. Check Angus Croll's solution.
The TL;DR version is below, but the article is a great discussion of the issue so you should read it if you have time.
Object.toType = function(obj) {
return ({}).toString.call(obj).match(/\s([a-z|A-Z]+)/)[1].toLowerCase();
}
// ... and usage:
Object.toType([1,2,3]); //"array" (all browsers)
// or to test...
var shouldBeAnArray = [1,2,3];
if(Object.toType(shouldBeAnArray) === 'array'){/* do stuff */};

If the only two kinds of values that could be passed to this function are a string or an array of strings, keep it simple and use a typeof check for the string possibility:
function someFunc(arg) {
var arr = (typeof arg == "string") ? [arg] : arg;
}

Here's my lazy approach:
if (Array.prototype.array_ === undefined) {
Array.prototype.array_ = true;
}
// ...
var test = [],
wat = {};
console.log(test.array_ === true); // true
console.log(wat.array_ === true); // false
I know it's sacrilege to "mess with" the prototype, but it appears to perform significantly better than the recommended toString method.
Note: A pitfall of this approach is that it wont work across iframe boundaries, but for my use case this is not an issue.

This function will turn almost anything into an array:
function arr(x) {
if(x === null || x === undefined) {
return [];
}
if(Array.isArray(x)) {
return x;
}
if(isString(x) || isNumber(x)) {
return [x];
}
if(x[Symbol.iterator] !== undefined || x.length !== undefined) {
return Array.from(x);
}
return [x];
}
function isString(x) {
return Object.prototype.toString.call(x) === "[object String]"
}
function isNumber(x) {
return Object.prototype.toString.call(x) === "[object Number]"
}
It uses some newer browser features so you may want to polyfill this for maximum support.
Examples:
> arr(null);
[]
> arr(undefined)
[]
> arr(3.14)
[ 3.14 ]
> arr(1/0)
[ Infinity ]
> gen = function*() { yield 1; yield 2; yield 3; }
[Function: gen]
> arr(gen())
[ 1, 2, 3 ]
> arr([4,5,6])
[ 4, 5, 6 ]
> arr("foo")
[ 'foo' ]
N.B. strings will be converted into an array with a single element instead of an array of chars. Delete the isString check if you would prefer it the other way around.
I've used Array.isArray here because it's the most robust and also simplest.

The following could be used if you know that your object doesn't have a concat method.
var arr = [];
if (typeof arr.concat === 'function') {
console.log("It's an array");
}

var a = [], b = {};
console.log(a.constructor.name == "Array");
console.log(b.constructor.name == "Object");

There is a nice example in Stoyan Stefanov's book JavaScript Patterns which is supposed to handle all possible problems as well as use the ECMAScript 5 method Array.isArray().
So here it is:
if (typeof Array.isArray === "undefined") {
Array.isArray = function (arg) {
return Object.prototype.toString.call(arg) === "[object Array]";
};
}
By the way, if you are using jQuery, you can use its method $.isArray().

You could use the isArray method, but I would prefer to check with:
Object.getPrototypeOf(yourvariable) === Array.prototype

function isArray(value) {
if (value) {
if (typeof value === 'object') {
return (Object.prototype.toString.call(value) == '[object Array]')
}
}
return false;
}
var ar = ["ff","tt"]
alert(isArray(ar))

A simple function for testing if an input value is an array is the following:
function isArray(value)
{
return Object.prototype.toString.call(value) === '[object Array]';
}
This works cross browser, and with older browsers. This is pulled from T.J. Crowders' blog post

You can try this:
var arr = []; (or) arr = new Array();
var obj = {}; (or) arr = new Object();
arr.constructor.prototype.hasOwnProperty('push') //true
obj.constructor.prototype.hasOwnProperty('push') // false

In your case you may use concat method of Array which can accept single objects as well as array (and even combined):
function myFunc(stringOrArray)
{
var arr = [].concat(stringOrArray);
console.log(arr);
arr.forEach(function(item, i)
{
console.log(i, "=", item);
})
}
myFunc("one string");
myFunc(["one string", "second", "third"]);
concat seems to be one of the oldest methods of Array (even IE 5.5 knows it well).

How to avoid 'cannot read property of undefined' errors?

In my code, I deal with an array that has some entries with many objects nested inside one another, where as some do not. It looks something like the following:
// where this array is hundreds of entries long, with a mix
// of the two examples given
var test = [{'a':{'b':{'c':"foo"}}}, {'a': "bar"}];
This is giving me problems because I need to iterate through the array at times, and the inconsistency is throwing me errors like so:
for (i=0; i<test.length; i++) {
// ok on i==0, but 'cannot read property of undefined' on i==1
console.log(a.b.c);
}
I am aware that I can say if(a.b){ console.log(a.b.c)}, but this is extraordinarily tedious in cases where there are up to 5 or 6 objects nested within one another. Is there any other (easier) way that I can have it ONLY do the console.log if it exists, but without throwing an error?

Update:
If you use JavaScript according to ECMAScript 2020 or later, see optional chaining.
TypeScript has added support for optional chaining in version 3.7.
// use it like this
obj?.a?.lot?.of?.properties
Solution for JavaScript before ECMASCript 2020 or TypeScript older than version 3.7:
A quick workaround is using a try/catch helper function with ES6 arrow function:
function getSafe(fn, defaultVal) {
try {
return fn();
} catch (e) {
return defaultVal;
}
}
// use it like this
console.log(getSafe(() => obj.a.lot.of.properties));
// or add an optional default value
console.log(getSafe(() => obj.a.lot.of.properties, 'nothing'));

What you are doing raises an exception (and rightfully so).
You can always do:
try{
window.a.b.c
}catch(e){
console.log("YO",e)
}
But I wouldn't, instead think of your use case.
Why are you accessing data, 6 levels nested that you are unfamiliar of? What use case justifies this?
Usually, you'd like to actually validate what sort of object you're dealing with.
Also, on a side note you should not use statements like if(a.b) because it will return false if a.b is 0 or even if it is "0". Instead check if a.b !== undefined

If I am understanding your question correctly, you want the safest way to determine if an object contains a property.
The easiest way is to use the in operator.
window.a = "aString";
//window should have 'a' property
//lets test if it exists
if ("a" in window){
//true
}
if ("b" in window){
//false
}
Of course you can nest this as deep as you want
if ("a" in window.b.c) { }
Not sure if this helps.

Try this. If a.b is undefined, it will leave the if statement without any exception.
if (a.b && a.b.c) {
console.log(a.b.c);
}

If you are using lodash, you could use their has function. It is similar to the native "in", but allows paths.
var testObject = {a: {b: {c: 'walrus'}}};
if(_.has(testObject, 'a.b.c')) {
//Safely access your walrus here
}

If you use Babel, you can already use the optional chaining syntax with #babel/plugin-proposal-optional-chaining Babel plugin. This would allow you to replace this:
console.log(a && a.b && a.b.c);
with this:
console.log(a?.b?.c);

If you have lodash you can use its .get method
_.get(a, 'b.c.d.e')
or give it a default value
_.get(a, 'b.c.d.e', default)

I use undefsafe religiously. It tests each level down into your object until it either gets the value you asked for, or it returns "undefined". But never errors.

This is a common issue when working with deep or complex json object, so I try to avoid try/catch or embedding multiple checks which would make the code unreadable, I usually use this little piece of code in all my procect to do the job.
/* ex: getProperty(myObj,'aze.xyz',0) // return myObj.aze.xyz safely
* accepts array for property names:
* getProperty(myObj,['aze','xyz'],{value: null})
*/
function getProperty(obj, props, defaultValue) {
var res, isvoid = function(x){return typeof x === "undefined" || x === null;}
if(!isvoid(obj)){
if(isvoid(props)) props = [];
if(typeof props === "string") props = props.trim().split(".");
if(props.constructor === Array){
res = props.length>1 ? getProperty(obj[props.shift()],props,defaultValue) : obj[props[0]];
}
}
return typeof res === "undefined" ? defaultValue: res;
}

I like Cao Shouguang's answer, but I am not fond of passing a function as parameter into the getSafe function each time I do the call. I have modified the getSafe function to accept simple parameters and pure ES5.
/**
* Safely get object properties.
* #param {*} prop The property of the object to retrieve
* #param {*} defaultVal The value returned if the property value does not exist
* #returns If property of object exists it is returned,
* else the default value is returned.
* #example
* var myObj = {a : {b : 'c'} };
* var value;
*
* value = getSafe(myObj.a.b,'No Value'); //returns c
* value = getSafe(myObj.a.x,'No Value'); //returns 'No Value'
*
* if (getSafe(myObj.a.x, false)){
* console.log('Found')
* } else {
* console.log('Not Found')
* }; //logs 'Not Found'
*
* if(value = getSafe(myObj.a.b, false)){
* console.log('New Value is', value); //logs 'New Value is c'
* }
*/
function getSafe(prop, defaultVal) {
return function(fn, defaultVal) {
try {
if (fn() === undefined) {
return defaultVal;
} else {
return fn();
}
} catch (e) {
return defaultVal;
}
}(function() {return prop}, defaultVal);
}

Lodash has a get method which allows for a default as an optional third parameter, as show below:
const myObject = {
has: 'some',
missing: {
vars: true
}
}
const path = 'missing.const.value';
const myValue = _.get(myObject, path, 'default');
console.log(myValue) // prints out default, which is specified above
<script src="https://cdnjs.cloudflare.com/ajax/libs/lodash.js/4.17.11/lodash.js"></script>

Imagine that we want to apply a series of functions to x if and only if x is non-null:
if (x !== null) x = a(x);
if (x !== null) x = b(x);
if (x !== null) x = c(x);
Now let's say that we need to do the same to y:
if (y !== null) y = a(y);
if (y !== null) y = b(y);
if (y !== null) y = c(y);
And the same to z:
if (z !== null) z = a(z);
if (z !== null) z = b(z);
if (z !== null) z = c(z);
As you can see without a proper abstraction, we'll end up duplicating code over and over again. Such an abstraction already exists: the Maybe monad.
The Maybe monad holds both a value and a computational context:
The monad keeps the value safe and applies functions to it.
The computational context is a null check before applying a function.
A naive implementation would look like this:
⚠️ This implementation is for illustration purpose only! This is not how it should be done and is wrong at many levels. However this should give you a better idea of what I am talking about.
As you can see nothing can break:
We apply a series of functions to our value
If at any point, the value becomes null (or undefined) we just don't apply any function anymore.
const abc = obj =>
Maybe
.of(obj)
.map(o => o.a)
.map(o => o.b)
.map(o => o.c)
.value;
const values = [
{},
{a: {}},
{a: {b: {}}},
{a: {b: {c: 42}}}
];
console.log(
values.map(abc)
);
<script>
function Maybe(x) {
this.value = x; //-> container for our value
}
Maybe.of = x => new Maybe(x);
Maybe.prototype.map = function (fn) {
if (this.value == null) { //-> computational context
return this;
}
return Maybe.of(fn(this.value));
};
</script>
Appendix 1
I cannot explain what monads are as this is not the purpose of this post and there are people out there better at this than I am. However as Eric Elliot said in hist blog post JavaScript Monads Made Simple:
Regardless of your skill level or understanding of category theory, using monads makes your code easier to work with. Failing to take advantage of monads may make your code harder to work with (e.g., callback hell, nested conditional branches, more verbosity).
Appendix 2
Here's how I'd solve your issue using the Maybe monad from monetjs
const prop = key => obj => Maybe.fromNull(obj[key]);
const abc = obj =>
Maybe
.fromNull(obj)
.flatMap(prop('a'))
.flatMap(prop('b'))
.flatMap(prop('c'))
.orSome('🌯')
const values = [
{},
{a: {}},
{a: {b: {}}},
{a: {b: {c: 42}}}
];
console.log(
values.map(abc)
);
<script src="https://www.unpkg.com/monet#0.9.0/dist/monet.js"></script>
<script>const {Maybe} = Monet;</script>

In str's answer, value 'undefined' will be returned instead of the set default value if the property is undefined. This sometimes can cause bugs. The following will make sure the defaultVal will always be returned when either the property or the object is undefined.
const temp = {};
console.log(getSafe(()=>temp.prop, '0'));
function getSafe(fn, defaultVal) {
try {
if (fn() === undefined || fn() === null) {
return defaultVal
} else {
return fn();
}
} catch (e) {
return defaultVal;
}
}

You can use optional chaining from the ECMAScript standart.
Like this:
a?.b?.c?.d?.func?.()

I answered this before and happened to be doing a similar check today. A simplification to check if a nested dotted property exists. You could modify this to return the value, or some default to accomplish your goal.
function containsProperty(instance, propertyName) {
// make an array of properties to walk through because propertyName can be nested
// ex "test.test2.test.test"
let walkArr = propertyName.indexOf('.') > 0 ? propertyName.split('.') : [propertyName];
// walk the tree - if any property does not exist then return false
for (let treeDepth = 0, maxDepth = walkArr.length; treeDepth < maxDepth; treeDepth++) {
// property does not exist
if (!Object.prototype.hasOwnProperty.call(instance, walkArr[treeDepth])) {
return false;
}
// does it exist - reassign the leaf
instance = instance[walkArr[treeDepth]];
}
// default
return true;
}
In your question you could do something like:
let test = [{'a':{'b':{'c':"foo"}}}, {'a': "bar"}];
containsProperty(test[0], 'a.b.c');

I usually use like this:
var x = object.any ? object.any.a : 'def';

You can avoid getting an error by giving a default value before getting the property
var test = [{'a':{'b':{'c':"foo"}}}, {'a': "bar"}];
for (i=0; i<test.length; i++) {
const obj = test[i]
// No error, just undefined, which is ok
console.log(((obj.a || {}).b || {}).c);
}
This works great with arrays too:
const entries = [{id: 1, name: 'Scarllet'}]
// Giving a default name when is empty
const name = (entries.find(v => v.id === 100) || []).name || 'no-name'
console.log(name)

Unrelated to the question's actual question, but might be useful for people coming to this question looking for answers.
Check your function parameters.
If you have a function like const x({ a }) => { }, and you call it without arguments x(); append = {} to the parameter: const x({ a } = {}) => { }.
What I had
I had a function like this:
const x = ({ a }) => console.log(a);
// This one works as expected
x({ a: 1 });
// This one errors out
x();
Which results in "Uncaught TypeError: Cannot destructure property 'a' of 'undefined' as it is undefined."
What I switched it to (now works).
const x = ({ a } = {}) => console.log(a);
// This one works as expected
x({ a: 1 });
// This now works too!
x();