me and my partner have been cracking our heads at this. we have to create a tree, that obviously has "children" below it.
all we need right now, is to loop over an object to find a certain value, if that value is not in that certain object, then go into its child property and look there.
basically what I'm asking is, how can you loop over nested objects until a certain value is found?
would super appreciate the perspective of a more experienced coder on this.
/// this is how one parent with a child tree looks like right now,
essentially if we presume that the child has another child in the children property,
how would we loop into that?
and maybe if that child also has a child, so on and so on...
Tree {
value: 'Parent',
children: [ Tree { value: 'Child', children: [] } ]
}
You can use recursive function to loop through objects:
const Tree = {
value: 'Parent',
children: [
{
value: 'Child1',
children: [
]
},
{
value: 'Child2',
children: [
{
value: 'Child2.1',
children: [
{
value: 'Child2.1.1',
children: [
]
},
{
value: 'Child2.1.2',
children: [
]
},
]
},
]
},
]
}
function findValue(obj, value)
{
if (obj.value == value)
return obj;
let ret = null;
for(let i = 0; i < obj.children.length; i++)
{
ret = findValue(obj.children[i], value);
if (ret)
break;
}
return ret;
}
console.log("Child1", findValue(Tree, "Child1"));
console.log("Child2.1", findValue(Tree, "Child2.1"));
console.log("Child3", findValue(Tree, "Child3"));
You can try this simple solution:
const myTree = {
value: 'Parent',
children: [
{
value: 'Child1',
children: []
},
{
value: 'Child2'
}
]
}
const isValueInTree = (tree, findValue) => {
if (tree.value === findValue) return true;
if (tree.children && tree.children.length !== 0) {
for (const branch of tree.children) {
const valuePresents = isValueInTree(branch, findValue);
if (valuePresents) return true;
}
}
return false;
}
console.log(isValueInTree(myTree, 'Child2'));
console.log(isValueInTree(myTree, 'Child'));
console.log(isValueInTree(myTree, 'Child1'));
See edit below
I wanted to try and create a tree from a list of paths and found this code on stackoverflow from another question and it seems to work fine but i would like to remove the empty children arrays instead of having them showing with zero items.
I tried counting r[name].result length and only pushing it if it greater than zero but i just end up with no children on any of the nodes.
let paths = ["About.vue","Categories/Index.vue","Categories/Demo.vue","Categories/Flavors.vue","Categories/Types/Index.vue","Categories/Types/Other.vue"];
let result = [];
let level = {result};
paths.forEach(path => {
path.split('/').reduce((r, name, i, a) => {
if(!r[name]) {
r[name] = {result: []};
r.result.push({name, children: r[name].result})
}
return r[name];
}, level)
})
console.log(result)
EDIT
I didnt want to ask directly for the purpose i am using it for but if it helps i am trying to create an array like this: (this is a copy paste of the config needed from ng-zorro cascader)
const options = [
{
value: 'zhejiang',
label: 'Zhejiang',
children: [
{
value: 'hangzhou',
label: 'Hangzhou',
children: [
{
value: 'xihu',
label: 'West Lake',
isLeaf: true
}
]
},
{
value: 'ningbo',
label: 'Ningbo',
isLeaf: true
}
]
},
{
value: 'jiangsu',
label: 'Jiangsu',
children: [
{
value: 'nanjing',
label: 'Nanjing',
children: [
{
value: 'zhonghuamen',
label: 'Zhong Hua Men',
isLeaf: true
}
]
}
]
}
];
from an array of flat fields like this:
let paths = ["About.vue","Categories/Index.vue","Categories/Demo.vue","Categories/Flavors.vue","Categories/Types/Index.vue","Categories/Types/Other.vue"];
I suggest to use a different approach.
This approach takes an object and not an array for reaching deeper levels and assigns an array only if the nested level is required.
let paths = ["About.vue", "Categories/Index.vue", "Categories/Demo.vue", "Categories/Flavors.vue", "Categories/Types/Index.vue", "Categories/Types/Other.vue"],
result = paths
.reduce((parent, path) => {
path.split('/').reduce((r, name, i, { length }) => {
let temp = (r.children ??= []).find(q => q.name === name);
if (!temp) r.children.push(temp = { name, ...(i + 1 === length && { isLeaf: true }) });
return temp;
}, parent);
return parent;
}, { children: [] })
.children;
console.log(result)
.as-console-wrapper { max-height: 100% !important; top: 0; }
I am building an object from a form that is currently rendered server side. I collect all the check boxes displayed in the image below and I am trying to sort them in a way that all the check boxes under each step (1, 2, 3 etc) is a single object based on the property parentNode.
Currently the document.querySelectorAll('.checkboxes') fetches all the checkboxes in following format.
var newObj = [
{
name: 'one',
parentNode: {
id: 'stepOne'
}
},
{
name: 'two',
parentNode: {
id: 'stepTwo'
}
},
{
name: 'three',
parentNode: {
id: 'stepOne'
}
},
]
The new object should be:
var newObj = {
stepOne: [
{
name: 'one',
parentNode: {
id: 'stepOne'
}
},
{
name: 'three',
parentNode: {
id: 'stepOne'
}
},
],
stepTwo: [
{
name: 'two',
parentNode: {
id: 'stepTwo'
}
},
]
}
Usually I do something like this:
let stepOne = function(step) {
return step.parentNode.getAttribute('id') === 'stepOne';
}
let stepTwo = function(step) {
return step.parentNode.getAttribute('id') === 'stepTwo';
}
let allTheStepOnes = fetchCheckBoxes.filter(stepOne);
But filter doesn't work on dom object and this seems inefficient as well.
Proper way of doing this is a forEach loop and using associative arrays like this:
let newObject = {};
originalObject.forEach((item)=>{
let step = item.parentNode.id
if (newObj[step] === undefined) {
newObj[step] = []
}
newObj[step].push(item)
})
Using reduce we can reduce your current array into the new structure.
return newObj.reduce(function(acc, item) {
If acc[item.parentNode.id] has been defined before, retrieve this. Otherwise set it to an empty array:
acc[item.parentNode.id] = (acc[item.parentNode.id] || [])
Add the item to the array and then return it:
acc[item.parentNode.id].push(item);
return acc;
We set the accumulator as {} to start with.
Snippet to show the workings.
var newObj = [{
name: 'one',
parentNode: {
id: 'stepOne'
}
}, {
name: 'two',
parentNode: {
id: 'stepTwo'
}
}, {
name: 'three',
parentNode: {
id: 'stepOne'
}
}, ];
var newOrder = function(prevList) {
return prevList.reduce(function(acc, item) {
acc[item.parentNode.id] = (acc[item.parentNode.id] || [])
acc[item.parentNode.id].push(item);
return acc;
}, {});
}
console.log(newOrder(newObj));
This function should do the trick
function mapObj(obj) {
var result = {};
for(var i = 0; i < obj.length; i++) {
var e = obj[i];
result[e.parentNode.id] = result[e.parentNode.id] || [];
result[e.parentNode.id].push(e);
}
return result;
}
See jsfiddle here: https://jsfiddle.net/remenyLx/2/
I have data that contains objects that each have an array of images. I want only the first image of each object.
var data1 = [
{
id: 1,
images: [
{ name: '1a' },
{ name: '1b' }
]
},
{
id: 2,
images: [
{ name: '2a' },
{ name: '2b' }
]
},
{
id: 3
},
{
id: 4,
images: []
}
];
var filtered = [];
var b = data1.forEach((element, index, array) => {
if(element.images && element.images.length)
filtered.push(element.images[0].name);
});
console.log(filtered);
The output needs to be flat:
['1a', '2a']
How can I make this prettier?
I'm not too familiar with JS map, reduce and filter and I think those would make my code more sensible; the forEach feels unnecessary.
First you can filter out elements without proper images property and then map it to new array:
const filtered = data1
.filter(e => e.images && e.images.length)
.map(e => e.images[0].name)
To do this in one loop you can use reduce function:
const filtered = data1.reduce((r, e) => {
if (e.images && e.images.length) {
r.push(e.images[0].name)
}
return r
}, [])
You can use reduce() to return this result.
var data1 = [{
id: 1,
images: [{
name: '1a'
}, {
name: '1b'
}]
}, {
id: 2,
images: [{
name: '2a'
}, {
name: '2b'
}]
}, {
id: 3
}, {
id: 4,
images: []
}];
var result = data1.reduce(function(r, e) {
if (e.hasOwnProperty('images') && e.images.length) r.push(e.images[0].name);
return r;
}, [])
console.log(result);
All answers are creating NEW arrays before projecting the final result : (filter and map creates a new array each) so basically it's creating twice.
Another approach is only to yield expected values :
Using iterator functions
function* foo(g)
{
for (let i = 0; i < g.length; i++)
{
if (g[i]['images'] && g[i]["images"].length)
yield g[i]['images'][0]["name"];
}
}
var iterator = foo(data1) ;
var result = iterator.next();
while (!result.done)
{
console.log(result.value)
result = iterator.next();
}
This will not create any additional array and only return the expected values !
However if you must return an array , rather than to do something with the actual values , then use other solutions suggested here.
https://jsfiddle.net/remenyLx/7/
I have and object literal that is essentially a tree that does not have a fixed number of levels. How can I go about searching the tree for a particualy node and then return that node when found in an effcient manner in javascript?
Essentially I have a tree like this and would like to find the node with the title 'randomNode_1'
var data = [
{
title: 'topNode',
children: [
{
title: 'node1',
children: [
{
title: 'randomNode_1'
},
{
title: 'node2',
children: [
{
title: 'randomNode_2',
children:[
{
title: 'node2',
children: [
{
title: 'randomNode_3',
}]
}
]
}]
}]
}
]
}];
Basing this answer off of #Ravindra's answer, but with true recursion.
function searchTree(element, matchingTitle){
if(element.title == matchingTitle){
return element;
}else if (element.children != null){
var i;
var result = null;
for(i=0; result == null && i < element.children.length; i++){
result = searchTree(element.children[i], matchingTitle);
}
return result;
}
return null;
}
Then you could call it:
var element = data[0];
var result = searchTree(element, 'randomNode_1');
Here's an iterative solution:
var stack = [], node, ii;
stack.push(root);
while (stack.length > 0) {
node = stack.pop();
if (node.title == 'randomNode_1') {
// Found it!
return node;
} else if (node.children && node.children.length) {
for (ii = 0; ii < node.children.length; ii += 1) {
stack.push(node.children[ii]);
}
}
}
// Didn't find it. Return null.
return null;
Here's an iterative function using the Stack approach, inspired by FishBasketGordo's answer but taking advantage of some ES2015 syntax to shorten things.
Since this question has already been viewed a lot of times, I've decided to update my answer to also provide a function with arguments that makes it more flexible:
function search (tree, value, key = 'id', reverse = false) {
const stack = [ tree[0] ]
while (stack.length) {
const node = stack[reverse ? 'pop' : 'shift']()
if (node[key] === value) return node
node.children && stack.push(...node.children)
}
return null
}
This way, it's now possible to pass the data tree itself, the desired value to search and also the property key which can have the desired value:
search(data, 'randomNode_2', 'title')
Finally, my original answer used Array.pop which lead to matching the last item in case of multiple matches. In fact, something that could be really confusing. Inspired by Superole comment, I've made it use Array.shift now, so the first in first out behavior is the default.
If you really want the old last in first out behavior, I've provided an additional arg reverse:
search(data, 'randomNode_2', 'title', true)
My answer is inspired from FishBasketGordo's iterativ answer. It's a little bit more complex but also much more flexible and you can have more than just one root node.
/**searchs through all arrays of the tree if the for a value from a property
* #param aTree : the tree array
* #param fCompair : This function will receive each node. It's upon you to define which
condition is necessary for the match. It must return true if the condition is matched. Example:
function(oNode){ if(oNode["Name"] === "AA") return true; }
* #param bGreedy? : us true to do not stop after the first match, default is false
* #return an array with references to the nodes for which fCompair was true; In case no node was found an empty array
* will be returned
*/
var _searchTree = function(aTree, fCompair, bGreedy){
var aInnerTree = []; // will contain the inner children
var oNode; // always the current node
var aReturnNodes = []; // the nodes array which will returned
// 1. loop through all root nodes so we don't touch the tree structure
for(keysTree in aTree) {
aInnerTree.push(aTree[keysTree]);
}
while(aInnerTree.length > 0) {
oNode = aInnerTree.pop();
// check current node
if( fCompair(oNode) ){
aReturnNodes.push(oNode);
if(!bGreedy){
return aReturnNodes;
}
} else { // if (node.children && node.children.length) {
// find other objects, 1. check all properties of the node if they are arrays
for(keysNode in oNode){
// true if the property is an array
if(oNode[keysNode] instanceof Array){
// 2. push all array object to aInnerTree to search in those later
for (var i = 0; i < oNode[keysNode].length; i++) {
aInnerTree.push(oNode[keysNode][i]);
}
}
}
}
}
return aReturnNodes; // someone was greedy
}
Finally you can use the function like this:
var foundNodes = _searchTree(data, function(oNode){ if(oNode["title"] === "randomNode_3") return true; }, false);
console.log("Node with title found: ");
console.log(foundNodes[0]);
And if you want to find all nodes with this title you can simply switch the bGreedy parameter:
var foundNodes = _searchTree(data, function(oNode){ if(oNode["title"] === "randomNode_3") return true; }, true);
console.log("NodeS with title found: ");
console.log(foundNodes);
FIND A NODE IN A TREE :
let say we have a tree like
let tree = [{
id: 1,
name: 'parent',
children: [
{
id: 2,
name: 'child_1'
},
{
id: 3,
name: 'child_2',
children: [
{
id: '4',
name: 'child_2_1',
children: []
},
{
id: '5',
name: 'child_2_2',
children: []
}
]
}
]
}];
function findNodeById(tree, id) {
let result = null
if (tree.id === id) {
return tree;
}
if (Array.isArray(tree.children) && tree.children.length > 0) {
tree.children.some((node) => {
result = findNodeById(node, id);
return result;
});
}
return result;}
You have to use recursion.
var currChild = data[0];
function searchTree(currChild, searchString){
if(currChild.title == searchString){
return currChild;
}else if (currChild.children != null){
for(i=0; i < currChild.children.length; i ++){
if (currChild.children[i].title ==searchString){
return currChild.children[i];
}else{
searchTree(currChild.children[i], searchString);
}
}
return null;
}
return null;
}
ES6+:
const deepSearch = (data, value, key = 'title', sub = 'children', tempObj = {}) => {
if (value && data) {
data.find((node) => {
if (node[key] == value) {
tempObj.found = node;
return node;
}
return deepSearch(node[sub], value, key, sub, tempObj);
});
if (tempObj.found) {
return tempObj.found;
}
}
return false;
};
const result = deepSearch(data, 'randomNode_1', 'title', 'children');
This function is universal and does search recursively.
It does not matter, if input tree is object(single root), or array of objects (many root objects). You can configure prop name that holds children array in tree objects.
// Searches items tree for object with specified prop with value
//
// #param {object} tree nodes tree with children items in nodesProp[] table, with one (object) or many (array of objects) roots
// #param {string} propNodes name of prop that holds child nodes array
// #param {string} prop name of searched node's prop
// #param {mixed} value value of searched node's prop
// #returns {object/null} returns first object that match supplied arguments (prop: value) or null if no matching object was found
function searchTree(tree, nodesProp, prop, value) {
var i, f = null; // iterator, found node
if (Array.isArray(tree)) { // if entry object is array objects, check each object
for (i = 0; i < tree.length; i++) {
f = searchTree(tree[i], nodesProp, prop, value);
if (f) { // if found matching object, return it.
return f;
}
}
} else if (typeof tree === 'object') { // standard tree node (one root)
if (tree[prop] !== undefined && tree[prop] === value) {
return tree; // found matching node
}
}
if (tree[nodesProp] !== undefined && tree[nodesProp].length > 0) { // if this is not maching node, search nodes, children (if prop exist and it is not empty)
return searchTree(tree[nodesProp], nodesProp, prop, value);
} else {
return null; // node does not match and it neither have children
}
}
I tested it localy and it works ok, but it somehow won't run on jsfiddle or jsbin...(recurency issues on those sites ??)
run code :
var data = [{
title: 'topNode',
children: [{
title: 'node1',
children: [{
title: 'randomNode_1'
}, {
title: 'node2',
children: [{
title: 'randomNode_2',
children: [{
title: 'node2',
children: [{
title: 'randomNode_3',
}]
}]
}]
}]
}]
}];
var r = searchTree(data, 'children', 'title', 'randomNode_1');
//var r = searchTree(data, 'children', 'title', 'node2'); // check it too
console.log(r);
It works in http://www.pythontutor.com/live.html#mode=edit (paste the code)
no BS version:
const find = (root, title) =>
root.title === title ?
root :
root.children?.reduce((result, n) => result || find(n, title), undefined)
This is basic recursion problem.
window.parser = function(searchParam, data) {
if(data.title != searchParam) {
returnData = window.parser(searchParam, children)
} else {
returnData = data;
}
return returnData;
}
here is a more complex option - it finds the first item in a tree-like node with providing (node, nodeChildrenKey, key/value pairs & optional additional key/value pairs)
const findInTree = (node, childrenKey, key, value, additionalKey?, additionalValue?) => {
let found = null;
if (additionalKey && additionalValue) {
found = node[childrenKey].find(x => x[key] === value && x[additionalKey] === additionalValue);
} else {
found = node[childrenKey].find(x => x[key] === value);
}
if (typeof(found) === 'undefined') {
for (const item of node[childrenKey]) {
if (typeof(found) === 'undefined' && item[childrenKey] && item[childrenKey].length > 0) {
found = findInTree(item, childrenKey, key, value, additionalKey, additionalValue);
}
}
}
return found;
};
export { findInTree };
Hope it helps someone.
A flexible recursive solution that will work for any tree
// predicate: (item) => boolean
// getChildren: (item) => treeNode[]
searchTree(predicate, getChildren, treeNode) {
function search(treeNode) {
if (!treeNode) {
return undefined;
}
for (let treeItem of treeNode) {
if (predicate(treeItem)) {
return treeItem;
}
const foundItem = search(getChildren(treeItem));
if (foundItem) {
return foundItem;
}
}
}
return search(treeNode);
}
find all parents of the element in the tree
let objects = [{
id: 'A',
name: 'ObjA',
children: [
{
id: 'A1',
name: 'ObjA1'
},
{
id: 'A2',
name: 'objA2',
children: [
{
id: 'A2-1',
name: 'objA2-1'
},
{
id: 'A2-2',
name: 'objA2-2'
}
]
}
]
},
{
id: 'B',
name: 'ObjB',
children: [
{
id: 'B1',
name: 'ObjB1'
}
]
}
];
let docs = [
{
object: {
id: 'A',
name: 'docA'
},
typedoc: {
id: 'TD1',
name: 'Typde Doc1'
}
},
{
object: {
id: 'A',
name: 'docA'
},
typedoc: {
id: 'TD2',
name: 'Typde Doc2'
}
},
{
object: {
id: 'A1',
name: 'docA1'
},
typedoc: {
id: 'TDx1',
name: 'Typde Doc x1'
}
},
{
object: {
id: 'A1',
name: 'docA1'
},
typedoc: {
id: 'TDx2',
name: 'Typde Doc x1'
}
},
{
object: {
id: 'A2',
name: 'docA2'
},
typedoc: {
id: 'TDx2',
name: 'Type de Doc x2'
}
},
{
object: {
id: 'A2-1',
name: 'docA2-1'
},
typedoc: {
id: 'TDx2-1',
name: 'Type de Docx2-1'
},
},
{
object: {
id: 'A2-2',
name: 'docA2-2'
},
typedoc: {
id: 'TDx2-2',
name: 'Type de Docx2-2'
},
},
{
object: {
id: 'B',
name: 'docB'
},
typedoc: {
id: 'TD1',
name: 'Typde Doc1'
}
},
{
object: {
id: 'B1',
name: 'docB1'
},
typedoc: {
id: 'TDx1',
name: 'Typde Doc x1'
}
}
];
function buildAllParents(doc, objects) {
for (let o = 0; o < objects.length; o++) {
let allParents = [];
let getAllParents = (o, eleFinded) => {
if (o.id === doc.object.id) {
doc.allParents = allParents;
eleFinded = true;
return { doc, eleFinded };
}
if (o.children) {
allParents.push(o.id);
for (let c = 0; c < o.children.length; c++) {
let { eleFinded, doc } = getAllParents(o.children[c], eleFinded);
if (eleFinded) {
return { eleFinded, doc };
} else {
continue;
}
}
}
return { eleFinded };
};
if (objects[o].id === doc.object.id) {
doc.allParents = [objects[o].id];
return doc;
} else if (objects[o].children) {
allParents.push(objects[o].id);
for (let c = 0; c < objects[o].children.length; c++) {
let eleFinded = null;`enter code here`
let res = getAllParents(objects[o].children[c], eleFinded);
if (res.eleFinded) {
return res.doc;
} else {
continue;
}
}
}
}
}
docs = docs.map(d => buildAllParents(d, objects`enter code here`))
This is an iterative breadth first search. It returns the first node that contains a child of a given name (nodeName) and a given value (nodeValue).
getParentNode(nodeName, nodeValue, rootNode) {
const queue= [ rootNode ]
while (queue.length) {
const node = queue.shift()
if (node[nodeName] === nodeValue) {
return node
} else if (node instanceof Object) {
const children = Object.values(node)
if (children.length) {
queue.push(...children)
}
}
}
return null
}
It would be used like this to solve the original question:
getParentNode('title', 'randomNode_1', data[0])
Enhancement of the code based on "Erick Petrucelli"
Remove the 'reverse' option
Add multi-root support
Add an option to control the visibility of 'children'
Typescript ready
Unit test ready
function searchTree(
tree: Record<string, any>[],
value: unknown,
key = 'value',
withChildren = false,
) {
let result = null;
if (!Array.isArray(tree)) return result;
for (let index = 0; index < tree.length; index += 1) {
const stack = [tree[index]];
while (stack.length) {
const node = stack.shift()!;
if (node[key] === value) {
result = node;
break;
}
if (node.children) {
stack.push(...node.children);
}
}
if (result) break;
}
if (withChildren !== true) {
delete result?.children;
}
return result;
}
And the tests can be found at: https://gist.github.com/aspirantzhang/a369aba7f84f26d57818ddef7d108682
Wrote another one based on my needs
condition is injected.
path of found branch is available
current path could be used in condition statement
could be used to map the tree items to another object
// if predicate returns true, the search is stopped
function traverse2(tree, predicate, path = "") {
if (predicate(tree, path)) return true;
for (const branch of tree.children ?? [])
if (traverse(branch, predicate, `${path ? path + "/" : ""}${branch.name}`))
return true;
}
example
let tree = {
name: "schools",
children: [
{
name: "farzanegan",
children: [
{
name: "classes",
children: [
{ name: "level1", children: [{ name: "A" }, { name: "B" }] },
{ name: "level2", children: [{ name: "C" }, { name: "D" }] },
],
},
],
},
{ name: "dastgheib", children: [{ name: "E" }, { name: "F" }] },
],
};
traverse(tree, (branch, path) => {
console.log("searching ", path);
if (branch.name === "C") {
console.log("found ", branch);
return true;
}
});
output
searching
searching farzanegan
searching farzanegan/classes
searching farzanegan/classes/level1
searching farzanegan/classes/level1/A
searching farzanegan/classes/level1/B
searching farzanegan/classes/level2
searching farzanegan/classes/level2/C
found { name: 'C' }
In 2022 use TypeScript and ES5
Just use basic recreation and built-in array method to loop over the array. Don't use Array.find() because this it will return the wrong node. Use Array.some() instead which allow you to break the loop.
interface iTree {
id: string;
children?: iTree[];
}
function findTreeNode(tree: iTree, id: string) {
let result: iTree | null = null;
if (tree.id === id) {
result = tree;
} else if (tree.children) {
tree.children.some((node) => {
result = findTreeNode(node, id);
return result; // break loop
});
}
return result;
}
const flattenTree = (data: any) => {
return _.reduce(
data,
(acc: any, item: any) => {
acc.push(item);
if (item.children) {
acc = acc.concat(flattenTree(item.children));
delete item.children;
}
return acc;
},
[]
);
};
An Approach to convert the nested tree into an object with depth 0.
We can convert the object in an object like this and can perform search more easily.
The following is working at my end:
function searchTree(data, value) {
if(data.title == value) {
return data;
}
if(data.children && data.children.length > 0) {
for(var i=0; i < data.children.length; i++) {
var node = traverseChildren(data.children[i], value);
if(node != null) {
return node;
}
}
}
return null;
}