Related
I faced a challenge where I needed to summarize an array of objects by the object's keys. I found a solution, but I can't shake off the feeling, that my approach is pretty naive:
const objArr = [
{ id: 1, val: "🍊" },
{ id: 1, val: "🍇" },
{ id: 1, val: "🍎" },
{ id: 2, val: "🥦" },
{ id: 2, val: "🌽" },
{ id: 2, val: "🌶" },
];
let tempArr = [];
let uniqueIdArr = [];
let sortedArr = [];
objArr.forEach((obj) => {
tempArr.push(obj.id);
uniqueIdArr = [...new Set(tempArr)];
});
uniqueIdArr.forEach((uniqueId) => {
let arr = [];
objArr.forEach((obj) => {
if (obj.id == uniqueId) {
arr.push(obj.val);
}
});
sortedArr.push({
id: uniqueId,
vals: arr,
});
});
console.log(sortedArr);
// Output: [{ id: 1, vals: [ '🍊', '🍇', '🍎' ] }, { id: 2, vals: [ '🥦', '🌽', '🌶' ] }]
Maybe there is something I don't know about JavaScript's array methods yet? Is this approach totally wrong? Is there another way, so that I could reduce the code and make it more elegant?
So many questions...
Any hint or explanation would be much appreciated. 🙈
Thanks in advance
J.
you can use Array.prototype.reduce to make your code bit shorter:
const objArr = [
{ id: 1, val: "🍊" },
{ id: 1, val: "🍇" },
{ id: 1, val: "🍎" },
{ id: 2, val: "🥦" },
{ id: 2, val: "🌽" },
{ id: 2, val: "🌶" },
];
let result = objArr.reduce((acc,e) => {
let idx = acc.findIndex(s => s.id === e.id)
if(idx > -1){
acc[idx].vals.push(e.val)
}
else{
acc.push({id:e.id,vals:[e.val]})
}
return acc
},[])
console.log(result)
Your ideas are good and explicit, but far from being optimal.
const objArr = [
{ id: 1, val: "🍊" },
{ id: 1, val: "🍇" },
{ id: 1, val: "🍎" },
{ id: 2, val: "🥦" },
{ id: 2, val: "🌽" },
{ id: 2, val: "🌶" },
];
const idValMap = new Map();
objArr.forEach(o=>{
let vals = idValMap.get(o.id);
if(!vals){
vals = [];
idValMap.set(o.id,vals);
}
vals.push(o.val);
});
console.log(Array.from(idValMap.entries()));
You can do most of it in just one loop. Take the key, check if you saw it already, if not initialize. That's it
This solution probably isn't much better but it is does use less code:
const objArr = [
{ id: 1, val: "🍊" },
{ id: 1, val: "🍇" },
{ id: 1, val: "🍎" },
{ id: 2, val: "🥦" },
{ id: 2, val: "🌽" },
{ id: 2, val: "🌶" },
];
let sortedArr = [];
objArr.forEach((item) => {
const exists = sortedArr.filter(i => i.id === item.id).length; // Check to see if we've already added an item with this ID
if (!exists) {
const matches = objArr.filter(i => i.id == item.id); // get all items with this ID
sortedArr.push({
id: item.id,
vals: matches.map(m => m.val) // We only care about the val property
});
}
});
See https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Array/map and https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Array/filter for informationa bout .map() and .filter() respectively
I've tried modifying some of the similar solutions on here but I keep getting stuck, I believe I have part of this figured out however, the main caveat is that:
Some of the objects have extra keys, which renders my object comparison logic useless.
I am trying to compare two arrays of objects. One array is the original array, and the other array contains the items I want deleted from the original array. However there's one extra issue in that the second array contains extra keys, so my comparison logic doesn't work.
An example would make this easier, let's say I have the following two arrays:
const originalArray = [{id: 1, name: "darnell"}, {id: 2, name: "funboi"},
{id: 3, name: "jackson5"}, {id: 4, name: "zelensky"}];
const itemsToBeRemoved = [{id: 2, name: "funboi", extraProperty: "something"},
{id: 4, name: "zelensky", extraProperty: "somethingelse"}];
after running the logic, my final output should be this array:
[{id: 1, name: "darnell"}, {id: 3, name: "jackson5"}]
And here's the current code / logic that I have, which compares but doesn't handle the extra keys. How should I handle this? Thank you in advance.
const prepareArray = (arr) => {
return arr.map((el) => {
if (typeof el === "object" && el !== null) {
return JSON.stringify(el);
} else {
return el;
}
});
};
const convertJSON = (arr) => {
return arr.map((el) => {
return JSON.parse(el);
});
};
const compareArrays = (arr1, arr2) => {
const currentArray = [...prepareArray(arr1)];
const deletedItems = [...prepareArray(arr2)];
const compared = currentArray.filter((el) => deletedItems.indexOf(el) === -1);
return convertJSON(compared);
};
How about using filter and some? You can extend the filter condition on select properties using &&.
const originalArray = [
{ id: 1, name: 'darnell' },
{ id: 2, name: 'funboi' },
{ id: 3, name: 'jackson5' },
{ id: 4, name: 'zelensky' },
];
const itemsToBeRemoved = [
{ id: 2, name: 'funboi', extraProperty: 'something' },
{ id: 4, name: 'zelensky', extraProperty: 'somethingelse' },
];
console.log(
originalArray.filter(item => !itemsToBeRemoved.some(itemToBeRemoved => itemToBeRemoved.id === item.id))
)
Or you can generalise it as well.
const originalArray = [
{ id: 1, name: 'darnell' },
{ id: 2, name: 'funboi' },
{ id: 3, name: 'jackson5' },
{ id: 4, name: 'zelensky' },
];
const itemsToBeRemoved = [
{ id: 2, name: 'funboi', extraProperty: 'something' },
{ id: 4, name: 'zelensky', extraProperty: 'somethingelse' },
];
function filterIfSubset(originalArray, itemsToBeRemoved) {
const filteredArray = [];
for (let i = 0; i < originalArray.length; i++) {
let isSubset = false;
for (let j = 0; j < itemsToBeRemoved.length; j++) {
// check if whole object is a subset of the object in itemsToBeRemoved
if (Object.keys(originalArray[i]).every(key => originalArray[i][key] === itemsToBeRemoved[j][key])) {
isSubset = true;
}
}
if (!isSubset) {
filteredArray.push(originalArray[i]);
}
}
return filteredArray;
}
console.log(filterIfSubset(originalArray, itemsToBeRemoved));
Another simpler variation of the second approach:
const originalArray = [
{ id: 1, name: 'darnell' },
{ id: 2, name: 'funboi' },
{ id: 3, name: 'jackson5' },
{ id: 4, name: 'zelensky' },
];
const itemsToBeRemoved = [
{ id: 2, name: 'funboi', extraProperty: 'something' },
{ id: 4, name: 'zelensky', extraProperty: 'somethingelse' },
];
const removeSubsetObjectsIfExists = (originalArray, itemsToBeRemoved) => {
return originalArray.filter(item => {
const isSubset = itemsToBeRemoved.some(itemToBeRemoved => {
return Object.keys(item).every(key => {
return item[key] === itemToBeRemoved[key];
});
});
return !isSubset;
});
}
console.log(removeSubsetObjectsIfExists(originalArray, itemsToBeRemoved));
The example below is a reusable function, the third parameter is the key to which you compare values from both arrays.
Details are commented in example
const arr=[{id:1,name:"darnell"},{id:2,name:"funboi"},{id:3,name:"jackson5"},{id:4,name:"zelensky"}],del=[{id:2,name:"funboi",extraProperty:"something"},{id:4,name:"zelensky",extraProperty:"somethingelse"}];
/** Compare arrayA vs. delArray by a given key's value.
--- ex. key = 'id'
**/
function deleteByKey(arrayA, delArray, key) {
/* Get an array of only the values of the given key from delArray
--- ex. delList = [1, 2, 3, 4]
*/
const delList = delArray.map(obj => obj[key]);
/* On every object of arrayA compare delList values vs
current object's key's value
--- ex. current obj[id] = 2
--- [1, 2, 3, 4].includes(obj[id])
Any match returns an empty array and non-matches are returned
in it's own array.
--- ex. ? [] : [obj]
The final return is a flattened array of the non-matching objects
*/
return arrayA.flatMap(obj => delList.includes(obj[key]) ? [] : [obj]);
};
console.log(deleteByKey(arr, del, 'id'));
let ff = [{ id: 1, name: 'darnell' }, { id: 2, name: 'funboi' },
{ id: 3, name: 'jackson5' },
{ id: 4, name: 'zelensky' }]
let cc = [{ id: 2, name: 'funboi', extraProperty: 'something' },
{ id: 4, name: 'zelensky', extraProperty: 'somethingelse' }]
let ar = []
let out = []
const result = ff.filter(function(i){
ar.push(i.id)
cc.forEach(function(k){
out.push(k.id)
})
if(!out.includes(i.id)){
// console.log(i.id, i)
return i
}
})
console.log(result)
I'm doing a complex filter, for which I have an initial list of objects with unique ids.
And a map with N properties with a list of corresponding object copies for each prop; and three states for each prop: idle: 0, show: 1, hide: 2.
For now I managed to do this with lodash's differenceBy and intersectionBy. My filter function takes in an array of objects and mutates the array, by checking and filtering the array with every map prop.
But concerning efficiency and growing number of complexity, should this kind of problem be solved differently?
For example:
If this filter is applied to a big array of hex colors (length 100, 1000 or more)
colors [1, 2, 3, ...1000]
And prop map has a growing number of props, like tags, by which a user can mark colors and show/hide them on filter. Or at some point new prop states will be added.
prop1 0, 1, 2, ...10
prop2 0, 1, 2, ...10
prop3 0, 1, 2, ...10
...
prop100 0, 1, 2, ...10
Should this kind of problem be solved via graph or matrix algorithms or some other method respectively? And, if yes, to what I should look into?
My code for optimisation and efficiency concerns:
const propMap = [
{ name: 'prop1', value: 0, items: [] },
{ name: 'prop2', value: 1, items: [ { id: 1}, { id: 2} ] },
{ name: 'propN', value: 2, items: [ { id: 2} ] },
];
const someArr = [
{ id: 1}, { id: 2}, { id: 3}, { id: 4},{ id: 5},
]
function filterByPropMap (arr) {
// Filter hidden from array
propMap.forEach(prop => {
if (prop.value === 2) {
arr = _.differenceBy(arr, prop.items, 'id');
}
});
// Filter intersecting objects to show
propMap.forEach(prop => {
if (prop.value === 1) {
arr = _.intersectionBy(arr, prop.items, 'id');
}
});
return [...arr];
}
console.log(filterByPropMap(someArr));
<script src="https://cdnjs.cloudflare.com/ajax/libs/lodash.js/4.17.20/lodash.min.js"></script>
In general, you very often don't need Lodash. Consider the following, which uses only vanilla ES6.
It returns { id: 1 } twice, because it looks like the deduplication is an unintended side effect of your code. At least you never do so explicitly.
const propMap = [
{ name: 'prop1', value: 0, items: [] },
{ name: 'prop2', value: 1, items: [ { id: 1}, { id: 2} ] },
{ name: 'propN', value: 2, items: [ { id: 2} ] },
];
const someArr = [
{ id: 1}, { id: 2}, { id: 3}, { id: 4},{ id: 1},
];
function filterByPropMap(arr) {
const hiddenItems = propMap
.filter(p => p.value === 2)
.map(p => p.items)
.flat();
const intersectingItems = propMap
.filter(p => p.value === 1)
.map(p => p.items)
.flat();
const isEqual = (a, b) => a.id === b.id;
return arr
.filter(v => !hiddenItems.some(h => isEqual(h, v)) &&
intersectingItems.some(i => isEqual(i, v)));
}
console.log(filterByPropMap(someArr));
How would I find all values by specific key in a deep nested object?
For example, if I have an object like this:
const myObj = {
id: 1,
children: [
{
id: 2,
children: [
{
id: 3
}
]
},
{
id: 4,
children: [
{
id: 5,
children: [
{
id: 6,
children: [
{
id: 7,
}
]
}
]
}
]
},
]
}
How would I get an array of all values throughout all nests of this obj by the key of id.
Note: children is a consistent name, and id's won't exist outside of a children object.
So from the obj, I would like to produce an array like this:
const idArray = [1, 2, 3, 4, 5, 6, 7]
This is a bit late but for anyone else finding this, here is a clean, generic recursive function:
function findAllByKey(obj, keyToFind) {
return Object.entries(obj)
.reduce((acc, [key, value]) => (key === keyToFind)
? acc.concat(value)
: (typeof value === 'object')
? acc.concat(findAllByKey(value, keyToFind))
: acc
, [])
}
// USAGE
findAllByKey(myObj, 'id')
You could make a recursive function like this:
idArray = []
function func(obj) {
idArray.push(obj.id)
if (!obj.children) {
return
}
obj.children.forEach(child => func(child))
}
Snippet for your sample:
const myObj = {
id: 1,
children: [{
id: 2,
children: [{
id: 3
}]
},
{
id: 4,
children: [{
id: 5,
children: [{
id: 6,
children: [{
id: 7,
}]
}]
}]
},
]
}
idArray = []
function func(obj) {
idArray.push(obj.id)
if (!obj.children) {
return
}
obj.children.forEach(child => func(child))
}
func(myObj)
console.log(idArray)
I found steve's answer to be most suited for my needs in extrapolating this out and creating a general recursive function. That said, I encountered issues when dealing with nulls and undefined values, so I extended the condition to accommodate for this. This approach uses:
Array.reduce() - It uses an accumulator function which appends the value's onto the result array. It also splits each object into it's key:value pair which allows you to take the following steps:
Have you've found the key? If so, add it to the array;
If not, have I found an object with values? If so, the key is possibly within there. Keep digging by calling the function on this object and append the result onto the result array; and
Finally, if this is not an object, return the result array unchanged.
Hope it helps!
const myObj = {
id: 1,
children: [{
id: 2,
children: [{
id: 3
}]
},
{
id: 4,
children: [{
id: 5,
children: [{
id: 6,
children: [{
id: 7,
}]
}]
}]
},
]
}
function findAllByKey(obj, keyToFind) {
return Object.entries(obj)
.reduce((acc, [key, value]) => (key === keyToFind)
? acc.concat(value)
: (typeof value === 'object' && value)
? acc.concat(findAllByKey(value, keyToFind))
: acc
, []) || [];
}
const ids = findAllByKey(myObj, 'id');
console.log(ids)
You can make a generic recursive function that works with any property and any object.
This uses Object.entries(), Object.keys(), Array.reduce(), Array.isArray(), Array.map() and Array.flat().
The stopping condition is when the object passed in is empty:
const myObj = {
id: 1,
anyProp: [{
id: 2,
thing: { a: 1, id: 10 },
children: [{ id: 3 }]
}, {
id: 4,
children: [{
id: 5,
children: [{
id: 6,
children: [{ id: 7 }]
}]
}]
}]
};
const getValues = prop => obj => {
if (!Object.keys(obj).length) { return []; }
return Object.entries(obj).reduce((acc, [key, val]) => {
if (key === prop) {
acc.push(val);
} else {
acc.push(Array.isArray(val) ? val.map(getIds).flat() : getIds(val));
}
return acc.flat();
}, []);
}
const getIds = getValues('id');
console.log(getIds(myObj));
Note: children is a consistent name, and id's wont exist outside
of a children object.
So from the obj, I would like to produce an array like this:
const idArray = [1, 2, 3, 4, 5, 6, 7]
Given that the question does not contain any restrictions on how the output is derived from the input and that the input is consistent, where the value of property "id" is a digit and id property is defined only within "children" property, save for case of the first "id" in the object, the input JavaScript plain object can be converted to a JSON string using JSON.stringify(), RegExp /"id":\d+/g matches the "id" property and one or more digit characters following the property name, which is then mapped to .match() the digit portion of the previous match using Regexp \d+ and convert the array value to a JavaScript number using addition operator +
const myObject = {"id":1,"children":[{"id":2,"children":[{"id":3}]},{"id":4,"children":[{"id":5,"children":[{"id":6,"children":[{"id":7}]}]}]}]};
let res = JSON.stringify(myObject).match(/"id":\d+/g).map(m => +m.match(/\d+/));
console.log(res);
JSON.stringify() replacer function can alternatively be used to .push() the value of every "id" property name within the object to an array
const myObject = {"id":1,"children":[{"id":2,"children":[{"id":3}]},{"id":4,"children":[{"id":5,"children":[{"id":6,"children":[{"id":7}]}]}]}]};
const getPropValues = (o, prop) =>
(res => (JSON.stringify(o, (key, value) =>
(key === prop && res.push(value), value)), res))([]);
let res = getPropValues(myObject, "id");
console.log(res);
Since the property values of the input to be matched are digits, all the JavaScript object can be converted to a string and RegExp \D can be used to replace all characters that are not digits, spread resulting string to array, and .map() digits to JavaScript numbers
let res = [...JSON.stringify(myObj).replace(/\D/g,"")].map(Number)
Using recursion.
const myObj = { id: 1, children: [ { id: 2, children: [ { id: 3 } ] }, { id: 4, children: [ { id: 5, children: [ { id: 6, children: [ { id: 7, } ] } ] } ] }, ]},
loop = (array, key, obj) => {
if (!obj.children) return;
obj.children.forEach(c => {
if (c[key]) array.push(c[key]); // is not present, skip!
loop(array, key, c);
});
},
arr = myObj["id"] ? [myObj["id"]] : [];
loop(arr, "id", myObj);
console.log(arr);
.as-console-wrapper { max-height: 100% !important; top: 0; }
You can make a recursive function with Object.entries like so:
const myObj = {
id: 1,
children: [{
id: 2,
children: [{
id: 3
}]
},
{
id: 4,
children: [{
id: 5,
children: [{
id: 6,
children: [{
id: 7,
}]
}]
}]
},
]
};
function findIds(obj) {
const entries = Object.entries(obj);
let result = entries.map(e => {
if (e[0] == "children") {
return e[1].map(child => findIds(child));
} else {
return e[1];
}
});
function flatten(arr, flat = []) {
for (let i = 0, length = arr.length; i < length; i++) {
const value = arr[i];
if (Array.isArray(value)) {
flatten(value, flat);
} else {
flat.push(value);
}
}
return flat;
}
return flatten(result);
}
var ids = findIds(myObj);
console.log(ids);
Flattening function from this answer
ES5 syntax:
var myObj = {
id: 1,
children: [{
id: 2,
children: [{
id: 3
}]
},
{
id: 4,
children: [{
id: 5,
children: [{
id: 6,
children: [{
id: 7,
}]
}]
}]
},
]
};
function findIds(obj) {
const entries = Object.entries(obj);
let result = entries.map(function(e) {
if (e[0] == "children") {
return e[1].map(function(child) {
return findIds(child)
});
} else {
return e[1];
}
});
function flatten(arr, flat = []) {
for (let i = 0, length = arr.length; i < length; i++) {
const value = arr[i];
if (Array.isArray(value)) {
flatten(value, flat);
} else {
flat.push(value);
}
}
return flat;
}
return flatten(result);
}
var ids = findIds(myObj);
console.log(ids);
let str = JSON.stringify(myObj);
let array = str.match(/\d+/g).map(v => v * 1);
console.log(array); // [1, 2, 3, 4, 5, 6, 7]
We use object-scan for a lot of our data processing needs now. It makes the code much more maintainable, but does take a moment to wrap your head around. Here is how you could use it to answer your question
// const objectScan = require('object-scan');
const find = (data, needle) => objectScan([needle], { rtn: 'value' })(data);
const myObj = { id: 1, children: [{ id: 2, children: [ { id: 3 } ] }, { id: 4, children: [ { id: 5, children: [ { id: 6, children: [ { id: 7 } ] } ] } ] }] };
console.log(find(myObj, '**.id'));
// => [ 7, 6, 5, 4, 3, 2, 1 ]
.as-console-wrapper {max-height: 100% !important; top: 0}
<script src="https://bundle.run/object-scan#13.7.1"></script>
Disclaimer: I'm the author of object-scan
import {flattenDeep} from 'lodash';
/**
* Extracts all values from an object (also nested objects)
* into a single array
*
* #param obj
* #returns
*
* #example
* const test = {
* alpha: 'foo',
* beta: {
* gamma: 'bar',
* lambda: 'baz'
* }
* }
*
* objectFlatten(test) // ['foo', 'bar', 'baz']
*/
export function objectFlatten(obj: {}) {
const result = [];
for (const prop in obj) {
const value = obj[prop];
if (typeof value === 'object') {
result.push(objectFlatten(value));
} else {
result.push(value);
}
}
return flattenDeep(result);
}
Below solution is generic which will return all values by matching nested keys as well e.g for below json object
{
"a":1,
"b":{
"a":{
"a":"red"
}
},
"c":{
"d":2
}
}
to find all values matching key "a" output should be return
[1,{a:"red"},"red"]
const findkey = (obj, key) => {
let arr = [];
if (isPrimitive(obj)) return obj;
for (let [k, val] of Object.entries(obj)) {
if (k === key) arr.push(val);
if (!isPrimitive(val)) arr = [...arr, ...findkey(val, key)];
}
return arr;
};
const isPrimitive = (val) => {
return val !== Object(val);
};
So, if I have two arrays...
const arr1 = [ { id: 1: newBid: true } ];
const arr2 = [ { id: 1, newBid: false }, { id: 2, newBid: false } ];
I want to wind up with an array that is like this
[ { id: 1, newBid: false }, { id: 2, newBid: false } ]
BUT... I want the { id: 1, newBid: true } to be from arr1 and not arr2
I was using Lodash uniqBy(arr1, arr2, ['id']), but it deletes the 1st occurance, not the 2nd
You should use lodash mergeWith function.
const arr1 = [{
id: 1,
newBid: true
}];
const arr2 = [{
id: 1,
newBid: false
}, {
id: 2,
newBid: false
}];
function customizer(firstValue, secondValue) {
if(firstValue)
return firstValue;
else
return secondValue;
}
console.log(_.mergeWith(arr1, arr2, customizer));
<script src="https://cdnjs.cloudflare.com/ajax/libs/lodash.js/4.17.10/lodash.js"></script>
I find using an object as a map to be the easiest way to deal with this kind of problems.
const arr1 = [ { id: 1, newBid: true } ];
const arr2 = [ { id: 1, newBid: false }, { id: 2, newBid: false } ];
const map = {};
function execute(array) {
for(let i = 0; i < array.length; i++) {
const item = array[i];
map[item.id] = item;
}
}
execute(arr1);
execute(arr2);
console.log(Object.values(map))