We Keep Coding

filtering list by multiple conditions

there is a list of users
filterData = [
{
"position":"lawyer",
"department_positions":[],
"group_positions":[
{"group":{"id":2,"code":"234","name":"group1"},"lead":false},
{"group":{"id":1,"code":"123","name":"group12"},"lead":true}
]
},
{
"position":"director",
"department_positions":[
{"department":{"id":3,"code":"333","name":"subDep"},"lead":false}
],
"group_positions":[
{"group":{"id":2,"code":"234","name":"group1"},"lead":false},
{"group":{"id":1,"code":"123","name":"group12"},"lead":true}
]
},
{
"position":"director",
"department_positions":[],
"group_positions":[]
}
]
and list of filters
categories = {
"position":["lawyer","director"],
"group_positions":["group1","group12"],
"department_positions":["generalDep", "subDep"]
}
It is necessary to filter users taking into account the fact that several filters can be selected at the same time. For example, i want to find user with position = "director" and AND group_positions = "group1" AND department_positions = "subDep"
my code doesn't allow filtering by multiple conditions. how can i fix it?
this.filter = this.filterData.filter(item => {
for (let key in this.categories) {
if (item[key].find(el =>
this.categories[key].includes(
el.group?.name || el.department?.name
)
)) {
return true
}
}
return false
})}

This is a good place to employ an es6 class to give behavior to the object being filtered. Augment each object to determine if it matches the "category" object.
(from the example data, this assumes the OP is looking for a "product of sums" match: for all of the category keys match at least one of the category values)
class FilterMe {
constructor(item) {
Object.assign(this, item);
}
namesForKey(key) {
switch (key) {
case 'position':
return [this.position]; // always answer an array
case 'group_positions':
return this.group_positions.map(gp => gp.group.name);
case 'department_positions':
return this.department_positions.map(dp => dp.department.name);
default:
return [];
}
}
// return true if a single filter key-value pair is matched
matchesFilterKeyValue(filterKey, filterOptions) {
const myNames = this.namesForKey(filterKey);
const matches = filterOptions.filter(e => myNames.includes(e));
return matches.length > 0;
}
// return true if all filter key-values pairs are matched
matchesFilter(filter) {
return Object.entries(filter).every(keyValue => {
return this.matchesFilterKeyValue(...keyValue);
})
}
}
const filterData = [{
"position": "lawyer",
"department_positions": [],
"group_positions": [{
"group": {
"id": 2,
"code": "234",
"name": "group1"
},
"lead": false
}, {
"group": {
"id": 1,
"code": "123",
"name": "group12"
},
"lead": true
}]
},
{
"position": "director",
"department_positions": [{
"department": {
"id": 3,
"code": "333",
"name": "subDep"
},
"lead": false
}],
"group_positions": [{
"group": {
"id": 2,
"code": "234",
"name": "group1"
},
"lead": false
}, {
"group": {
"id": 1,
"code": "123",
"name": "group12"
},
"lead": true
}]
},
{
"position": "director",
"department_positions": [],
"group_positions": []
}
]
const categories = {
"position": ["lawyer", "director"],
"group_positions": ["group1", "group12"],
"department_positions": ["generalDep", "subDep"]
}
// convert the filterData to the objects and test them...
let objects = filterData.map(d => new FilterMe(d));
let matches = objects.filter(o => o.matchesFilter(categories))
console.log(matches)

You can try something like this:
let filtered = example.filter(item => {
let valid = false
if (item.includes('something')) {
valid = true
}
if (!valid) {
// check second condition
}
return valid
})
Use a temporary placeholder so you don't immediately have to return true/false.

Counting multiple json inputs js

I get an input like this:
input 1:
{
"name": "Ben",
"description": "Ben",
"attributes": [
{
"type": "Background",
"value": "Default"
},
{
"type": "Hair-color",
"value": "Brown"
}
]
}
input 2
{
"name": "Ice",
"description": "Ice",
"attributes": [
{
"type": "Background",
"value": "Green"
},
{
"type": "Hair-color",
"value": "White"
}
]
}
input 3
{
"name": "Itay",
"description": "Itay",
"attributes": [
{
"type": "Background",
"value": "Default"
},
{
"type": "Hair-color",
"value": "Brown"
}
]
}
What I want to do is count the amount of each type of background and each type of hair-color appearing.
(These are sample examples and in reality there are more types and different values)
Let's say in these examples we have 2 objects that have a background as default then I want to have a count of that like so:
export interface TraitCount {
value: string,
count: number
}
export interface CountOfEachAttribute {
trait_type: string,
trait_count: traitCount[] | null,
total_variations: number
}
I want the most effective code because there are other aspects to the code, in addition it will run on 5-10k queries not just three, so needs
to run in good times too :D
(It's similar to my other question done with python but now I need it in js also)
Atm it's something like this:
(Apart of a much bigger code so keep that in mind)
setInitalCountOfAllAttribute( state, { payload }: PayloadAction<CountOfEachAttribute[] | null> ) {
if (payload === null) {
state.countOfAllAttribute = null;
} else {
state.countOfAllAttribute = payload;
}
},
setCountOfAllAttribute(state, { payload }: PayloadAction<Attribute>) {
if (state.countOfAllAttribute !== null) {
state.countOfAllAttribute.map(
(countOfEachAttribute: CountOfEachAttribute) => {
// Find the trait type
if (countOfEachAttribute.trait_type === payload.trait_type) {
// initiate the trait count array to store all the trait values and add first trait value
if (countOfEachAttribute.trait_count === null) {
const new_trait_count = { value: payload.value, count: 1 };
countOfEachAttribute.trait_count = [new_trait_count];
countOfEachAttribute.total_variations++;
}
// Trait array already existed.
else {
// Check if value already present or not
const checkValue = (obj: any) => obj.value === String(payload.value);
const isPresent = countOfEachAttribute.trait_count.some(checkValue)
const isPresent2 = countOfEachAttribute.trait_count.find((elem: any) => elem.value === String(payload.value))
// Value matched, increase its count by one
if (isPresent2) {
countOfEachAttribute.trait_count &&
countOfEachAttribute.trait_count.map((trait) => {
if (trait.value === payload.value) {
trait.count++;
}
});
}
// Value doesn't match, add a new entry and increase the count of variations by one
else {
const new_trait_count = { value: payload.value, count: 1 };
countOfEachAttribute.trait_count = [
...countOfEachAttribute.trait_count,
new_trait_count,
];
countOfEachAttribute.total_variations++;
}
}
}
}
);
}
},

You can merge all arrays and use Array.reduce.
const input1 = {
"name": "Ben",
"description": "Ben",
"attributes": [{
"type": "Background",
"value": "Default"
},
{
"type": "Hair-color",
"value": "Brown"
}
]
}
const input2 = {
"name": "Ice",
"description": "Ice",
"attributes": [{
"type": "Background",
"value": "Green"
},
{
"type": "Hair-color",
"value": "White"
}
]
}
const input3 = {
"name": "Itay",
"description": "Itay",
"attributes": [{
"type": "Background",
"value": "Default"
},
{
"type": "Hair-color",
"value": "Brown"
}
]
}
const mergedInput = [input1, input2, input3];
const result = mergedInput.reduce((acc, item) => {
item.attributes.forEach(attrItem => {
const existType = acc.find(e => e.trait_type == attrItem.type);
if (existType) {
var existAttr = existType.trait_count.find(e => e.value == attrItem.value);
if (existAttr) {
existAttr.count++;
} else {
existType.trait_count.push({
value: attrItem.value,
count: 1
});
existType.total_variations++;
}
} else {
acc.push({
trait_type: attrItem.type,
trait_count: [{
value: attrItem.value,
count: 1
}],
total_variations: 1
})
}
});
return acc;
}, []);
console.log(result);

I suggest instead of creating an array for trait_count to make it an object so you don't have to iterate over it whenever you are adding a new attribute. In the snippet below I'm using the value of the attribute as a sort of hash that allows the access to the given property without having to call the Array.prototype.find function
const input1 = {"name":"Ben","description":"Ben","attributes":[{"type":"Background","value":"Default"},{"type":"Hair-color","value":"Brown"}]};
const input2 = {"name":"Ice","description":"Ice","attributes":[{"type":"Background","value":"Green"},{"type":"Hair-color","value":"White"}]};
const input3 = {"name":"Itay","description":"Itay","attributes":[{"type":"Background","value":"Default"},{"type":"Hair-color","value":"Brown"}]};
function countAtributes(input, totalCounts={}) {
input.attributes.forEach((attribute) => {
if (!totalCounts[attribute.type])
totalCounts[attribute.type] = {trait_type: attribute.type, trait_count: {}, total_variations: 0};
if (!totalCounts[attribute.type].trait_count[attribute.value]) {
totalCounts[attribute.type].trait_count[attribute.value] = {value: attribute.value, count: 1};
totalCounts[attribute.type].total_variations+=1;
}
else totalCounts[attribute.type].trait_count[attribute.value].count +=1;
})
}
const totalCounts = {};
countAtributes(input1, totalCounts);
countAtributes(input2, totalCounts);
countAtributes(input3, totalCounts);
console.log(totalCounts);
It could be turned into the array afterwards with Object.values if necessary
I believe it is a much better approach to what you had before as you don't have to iterate over the tables of trait_counts. In theory it should significantly reduce the time taken. Iterating over the array and checking a condition each time is much slower than key lookup in Javascript object

Building multi-level aggregation queries using elastic-builder.js

I am trying to get my hands on javascript and elasticsearch and I was trying to create queries using the elastic-builder javascript lib.
I might be missing something which I am trying to figure out but unfortunately I am unable to.
Problem: I am trying to create multilevel aggregation like below,
"aggs": {
"1": {
"date_histogram": {
"field": "f1",
"calendar_interval": "1D"
},
"aggs": {
"2": {
"date_histogram": {
"field": "f2",
"calendar_interval": "1D"
},
"aggs": {
"3": {
"date_histogram": {
"field": "f3",
"calendar_interval": "1D"
}
}
}
}
}
}
But what I get instead is this:
"aggs": {
"1": {
"date_histogram": {
"field": "f1",
"calendar_interval": "1D"
},
"aggs": {
"2": {
"date_histogram": {
"field": "f2",
"calendar_interval": "1D"
}
},
"3": {
"date_histogram": {
"field": "f3",
"calendar_interval": "1D"
}
}
}
}
The current output I get has two aggregations nested in one. I am trying to build it using an array with aggregations defined in it.
The code I used is below:
let a = [
esb.dateHistogramAggregation('1', "d[key]['field']").calendarInterval('1D'),
esb.dateHistogramAggregation('2', "d[key]['field']").calendarInterval("1D"),
esb.dateHistogramAggregation('3', "d[key]['field']").calendarInterval("1D")
];
let m = null;
for(i=0;i<a.length;i++) {
if(i === 0) {
m = a[i]
} else {
m.agg(a[i])
}
}
//m = esb.dateHistogramAggregation('1', "d[key]['field']").calendarInterval('1D')
//m = m.agg(esb.dateHistogramAggregation('2', "d[key]['field']").calendarInterval("1D").agg(esb.dateHistogramAggregation('3', "d[key]['field']").calendarInterval("1D")))
esb.requestBodySearch()
.query(
esb.boolQuery()
.must(esb.matchQuery('message', 'this is a test'))
.filter(esb.termQuery('user', 'kimchy'))
.filter(esb.termQuery('user', 'herald'))
.should(esb.termQuery('user', 'johnny'))
.mustNot(esb.termQuery('user', 'cassie'))
)
.agg(esb.termsAggregation('user_terms', 'user').agg(esb.termsAggregation('user_terms', 'user').agg(esb.termsAggregation('user_terms', 'user'))))
.agg(m);
The lines commented in the code will output the result I'm expecting. What am I doing wrong?

You can turn the array into a group of sub-aggregations like so:
let a = [
esb.dateHistogramAggregation('1', "d[key]['field']").calendarInterval('1D'),
esb.dateHistogramAggregation('2', "d[key]['field']").calendarInterval("1D"),
esb.dateHistogramAggregation('3', "d[key]['field']").calendarInterval("1D")
];
const reqBody = esb.requestBodySearch()
.agg(
a[0].agg(
a[1].agg(
a[2]
)
)
);

I solved it like below. I am not sure this is right way. But someone can correct me if I am wrong.
let temp = null;
for (i = a.length - 1; i >= 0; i--) {
if (i === a.length - 1) {
temp = a[i];
} else {
temp = a[i].agg(temp)
}
}

How to convert Json into associative array or key value array

I have following Json which i need to insert into a table.
I want to convert each student detail into a row.
Because if i loop through the rows as per the existing structure i am reading one column as a row.
var json {
"Students":[
{
"name":{
"value":"Allan"
},
"number":{
"value":"123"
}
},
{
"name":{
"value":"Frank"
},
"number":{
"value":"456"
}
}
]
}
Ideally i want to the above as
{ "name": "Allan", "number": 123};
{ "name": "Frank", "number": 456};
I am looping through the Json as below
var objectKeys = Object.keys(json);
for (var key in objectKeys)
{
var student = json.Students;
for (var i = 0; i < student .length; i++) {
for (var column in json.Students[i]) {
window.print(column);
window.print(json.Students[i][column].value);
}
}
}
NOTE: No JQuery, want to achieve the above through normal Javascript.

If you want to transform the data, you can use Array.map
var json = {"Students":[{"name":{"value":"Allan"},"number":{"value":"123"}},{"name":{"value":"Frank"},"number":{"value":"456"}}]};
let result = json.Students.map(o => ({
name: o.name.value,
number: o.number.value
}));
console.log(result);
If you want to access the data, you can use Array.forEach
var json = {"Students":[{"name":{"value":"Allan"},"number":{"value":"123"}},{"name":{"value":"Frank"},"number":{"value":"456"}}]};
json.Students.forEach(o => console.log({name: o.name.value, number: o.number.value}));

var json = {
"Students":[
{
"name":{
"value":"Allan"
},
"number":{
"value":"123"
}
},
{
"name":{
"value":"Frank"
},
"number":{
"value":"456"
}
}
]
}
var studentData = JSON.stringify(json.Students);
var convertedData = JSON.parse(studentData.replace(/\{\"value\"\:/g,"").replace(/\}\,\"number/g,',"number').replace(/\"\}\}/g,'"}'));
Try this :)

No map or reduce. Just classic Javascript.
var json = {
"Students": [{
"name": {
"value": "Allan"
},
"number": {
"value": "123"
}
},
{
"name": {
"value": "Frank"
},
"number": {
"value": "456"
}
}
]
};
for (var student of json["Students"]) {
console.log(student); //your logic goes here.
}

Loop through object and replace values

I have an object which at some points is four levels deep, however I want a function that will cope should more levels be introduced. I'm trying to write a function that will replaced elements such that <span class="ajax-parent1-parent2-parent3-value"></span> will be replaced with parent1.parent2.parent3.value.
The issue is that the depth is variable, so I could have something like <span class="ajax-parent1-value"></span> to be replaced with parent1.value.
Finally, it's not always the text to be replaced. Optionally, data-attr can specify an attribute to be used instead (through element.attr(<data-attr>, <value>)).
Currently, I'm iterating manually, however it isn't very clean so I was wondering if there is a better way to do it. This also doesn't work for greater than two levels deep.
function render(data) {
$.each(data, function(parent, value) {
$.each(value, function(element, value) {
$el = $('.ajax-' + parent + '-' + element);
$.each($el, function(key, el) {
if ($(el).data('attr')) {
$(el).attr($(el).data('attr'), value);
} else {
$(el).text(value);
}
}
});
});
}
Example object:
{
"profile": {
"username": "johnd",
"bio": "Some example data about John",
"website": "http://john.com",
"profile_picture": "http://john.com/me.jpg",
"full_name": "John Doe",
"counts": {
"media": 13,
"followed_by": 26,
"follows": 49
},
"id": "16"
},
"dashboard": {
"script": {
"tags": ["media"],
"stats": {
"liked": 0,
"lastrun": "never",
"duration": 0
},
"status": {
"code": 0,
"slug": "disabled",
"human": "Disabled",
"message": "Not running."
}
},
"account": {
"plan": "free",
"created": 1419261005373,
"updated": 1419261005373
}
},
"serverInformation": {
"serverName": "Johns API",
"apiVersion": "0.0.1",
"requestDuration": 22,
"currentTime": 1419262805646
},
"requesterInformation": {
"id": "redacted",
"fingerprint": "redacted",
"remoteIP": "redacted",
"receivedParams": {
"action": "getDashboard",
"apiVersion": 1
}
}
}

Here is the solution I wrote:
function iterate(obj, stack) {
for (var property in obj) {
if (obj.hasOwnProperty(property)) {
if (typeof obj[property] == "object") {
iterate(obj[property], stack + '-' + property);
} else {
$group = $('.ajax' + stack + '-' + property);
$.each($group, function(key, element) {
if ($(element).data('attr')) {
$(element).attr($(element).data('attr'), obj[property]);
} else {
$(element).text(obj[property]);
}
});
}
}
}
}

Why don't you start from the HTML, so you only access the properties you actually want to render?
That way you can keep it quite simple (also note that this removes the need to nest HTML spans in the same order/depth as the data object, you can just place any HTML node anywhere. Just make sure you don't use class/node names more then once.
function parseData(data) {
var $container = $('.ajax');
$container.find("[class^='ajax-']").each(function(i, el) {
var $el = $(el);
if ($el.children().length === 0)
{
var nodes = $el.attr('class').split('-');
nodes.shift();
var node = data;
for (var i = 0; i < nodes.length; i++) {
node = node[nodes[i]];
if (typeof(node) == "undefined") {
break;
}
}
if ($el.data('attr'))
{
$el.attr($el.data('attr'), node);
}
else
{
$el.text(node);
}
}
});
}
Fiddle here:
http://jsfiddle.net/ckcduLhn/5/