View modules return the UI configuration of the components. You can put data loading logic in views, which is relevant for prototyping and considered bad practice for real-life apps.
Good practice is to separate data loading logic from UI configuration and put it in separate modules - models. A model stores all functionality related to some data entity. This division of UI and data is an advantage because if the data changes, all you have to do is change the data model. There is no need to modify all the views that use the data.
In the data model, you can define:
- a DataCollection that will load the data and manage saving;
- a method that will fetch the data from backend.
There are several types of models in Webix Jet:
- shared models
- dynamic models
- remote models
- services for data
- models with webix.remote
For relatively small data used in many components, you can load data into Webix DataCollection. This way, you make a single request to the server, store the data on the client side and sync it with the necessary views.
You can export either the whole collection or an accessor function that will return it.
// models/records.js
export const records = new webix.DataCollection({
url:"data.php"
});
//or
const records = new webix.DataCollection({
url:"data.php"
});
export function getRecords(){ return records; };A collection will handle save operations as well. You need to provide a save URL for a collection, so that it can send all updates made from views to the server:
// models/records.js
export const records = new webix.DataCollection({
url:"data.php",
save:"data.php"
});To use the data in a component, you can parse it. You must parse data in init(), not in config() (leave config() for the UI). Have a look at the example with a datatable:
// views/data.js
import {JetView} from "webix-jet";
import {records} from "models/records";
export default class DataView extends JetView{
config: () => {
view:"datatable", autoConfig:true, editable:true
}
init(view){
view.parse(records);
}
}All the changes made in the datatable are saved to the server.
You can use models to load options of select boxes, e.g.:
// views/start.js
import {options} from "models/options";
import {JetView} from "webix-jet";
export default class StartView extends JetView {
config(){
return {
view:"form", elements:[
{ view:"combo", options:{ data:options }}
]
};
}
});And for options of Datatable select editors:
// views/data.js
import {JetView} from "webix-jet";
import {options} from "models/options";
export default class DataView extends JetView{
config(){
return {
view:"datatable", editable:true, columns:[
{ id:"categoryId", editor:"richselect", collection:options }
]
}
}
}You can also sync a data component inside a Jet view with a DataCollection. Let's sync the datatable with records:
// views/data.js
import {JetView} from "webix-jet";
import {records} from "models/records";
export default class DataView extends JetView{
config: () => {
view:"datatable", autoConfig:true, editable:true
}
init(view){
view.sync(records);
}
removeRecord(id){
records.remove(id);
}
}Mind that if you synced a data component to a DataCollection, you must perform add/remove operations on the master collection, while the synced view will reflect these changes automatically. Slave views can only update the master.
To return several types of data and distribute data chunks to different views, a shared data transport can be used.
A shared data model can look like this:
//models/shareddata.js
const data = webix.ajax("some.json").then(a => a.json());
export function sharedData(name){
return data.then(a => {
switch (name){
case "grid":
return a.grid;
default:
return [];
}
});
}The model communicates with the shared data feed and can provide different data chunks for different views, for example, specific data for a grid and common data for other components.
Each component must have its own model as well. For example, this is the data model for a grid:
//models/griddata.js
import {sharedData} from "models/shareddata";
const gridData = new webix.DataCollection();
gridData.parse(sharedData("grid"));
export gridData;A component expects that sharedData will return an array of objects or a promise of an array of objects.
If you are returning data in a different format, you can add a data transformation for the promise chain:
//models/griddata.js
import {sharedData} from "models/shareddata";
const gridData = new webix.DataCollection();
let data = sharedData("grid");
data.then(function(d){
gridData.parse(d.value);
});
export gridData;A grid can use this model like this:
// views/grid.js
import {JetView} from "webix-jet";
import {gridData} from "models/griddata";
export default class GridView extends JetView{
config(){
return { view:"dataview", localId:"grid", autoConfig:true };
}
init(){
// for grids
this.$$("grid").sync(gridData);
}
}With forms, you should use a different approach:
// views/form.js
import {JetView} from "webix-jet";
import {gridData} from "models/griddata";
export default class FormView extends JetView {
config(){
return { view:"form", localId:"form", elements:[...] };
}
init(){
// for form
this.$$("form").setValues(gridData.getItem(id));
}
}where id is the ID of the record from DataCollection that you want to load into the form.
Note: if you have a single record in DataCollection, you can use the DataRecord object instead:
// models/data.js
import {sharedData} from "models/shareddata";
export const data = new webix.DataRecord({});
let temp = sharedData("grid");
temp.then(function(d){
data.setValues(d.value);
});// views/form.js
import {JetView} from "webix-jet";
import {data} from "models/data";
export default class FormView extends JetView {
config(){
return { view:"form", localId:"form", elements:[...] };
}
init(){
// for form
this.$$("form").setValues(data);
}
}or just a plain JS object:
import {sharedData} from "models/shareddata";
export const data = {};
export const ready = sharedData("grid").then(d => webix.extend(data,d) );// views/form.js
import {JetView} from "webix-jet";
import {ready, data} from "models/data";
export default class FormView extends JetView {
config(){
return { view:"form", localId:"form", elements:[...] };
}
init(){
ready.then( _ => this.$$("form").setValues(data) );
}
}As browser resources are limited, big collections should not be stored on the client side. Dynamic model is for big data (less than 10K records) used only once that must not be cached.
The data can be loaded from a server directly with an AJAX request:
// models/records.js
export function getData(){
return webix.ajax("data.php");
}It can be a service, a script, a function, etc. and it should return either a data object/array or a promise that will be resolved with the needed data.
To parse data, pass the return value of getData to view.parse:
// views/data.js
import {JetView} from "webix-jet";
import {getData} from "../models/records";
export default class DataView extends JetView{
config: () => {
view:"datatable", autoConfig:true
}
init(view){
view.parse(getData());
}
}You can also load data from local storage:
// models/records.js
export function getData(){
return webix.storage.local.get("data");
}To save data, you must provide a pattern for saving. Let's add one more function to the records model:
// models/records.js
export function getData(){
return webix.ajax("data.php");
};
export function saveData(id, operation, data){
if (operation is "update") //"add", "delete"
return webix.ajax().post("data.php", data);
if (operation =="add")
// ...
};The saveData() function provided as the save property of a data widget, enables DataProcessor, which will call this function for all additions, removals and updates. Let's define a way to save data in init() of a view:
// views/data.js
import {JetView} from "webix-jet";
import {getData, saveData} from "models/records";
export default class DataView extends JetView{
config: () => {
view:"datatable", autoConfig:true, editable:true
}
init(view) {
view.parse(getData());
view.define("save", saveData);
}
}Loading and saving data can also be in config of the view module, if needed:
// views/data.js
import {JetView} from "webix-jet";
import {getData, saveData} from "models/records";
export default class DataView extends JetView{
config: () => {
view:"datatable", autoConfig:true,
url: getData,
save: saveData
}
}In case data is really huge, you can drop the whole concept of separating data from UI and rely on the dynamic loading of Webix components (Datatable, Dataview, List, Tree). Data will be loaded in portions when needed. For that, load data in the view configuration. You can do it with the url property. For saving data, use the save property.
// views/data.js
import {JetView} from "webix-jet";
export default class DataView extends JetView{
config() => {
view:"datatable", autoConfig:true,
url:"data.php",
save:"data.php"
}
}Data can also be dynamically loaded with the load method:
// views/data.js
import {JetView} from "webix-jet";
export default class DataView extends JetView{
config(){
const ui = { view:"datatable", autoConfig:true };
ui.load("data.php");
return ui;
}
}Webix dynamic loading pattern allows loading data in portions, provided that your backend returns the response as: { data:[], total_count:1000, pos:100}. When the component is scrolled (or tree branches are opened) the component itself can send the request to the server for a new portion of data using the defined data url.
Let's recap main differences of the three ways to load and save data:
| Model | Data Size | Usage |
|---|---|---|
| Shared | Small | Many times |
| Dynamic | Big | Once |
| Remote | Huge | Handy for prototyping |
You can use services instead of models as data sources. Suppose there is a list of customers and a grid that displays records on a selected customer. Here is how you can use a service that returns the ID of a selected list item:
const id = service.getSelected();
const data = records.getData(id);A better and shorter way is:
const data = service.getNomenclature();You can use webix.remote instead of sending AJAX requests. You must have a server-side script with a class for getting the data:
//api.php
class Nomencl {
public function getData(id) { /*..*/ }
}
$api->setClass("nm", new Nomencl());The class with all its methods is used by webix.remote in a model:
// models/nomencl.js
export default webix.remote.nm;webix.remote can also be used to create a DataCollection:
// models/nomencl.js
const data = new DataCollection({
url: webix.remote.nm.getData
})Models with webix.remote can be imported and parsed it into views:
// views/data.js
import records from "models/nomencl";
...
init(view){
view.parse(records.getData(id));
}webix.remote is better then an AJAX request for several reasons:
1. You don't need to serialize data after loading.
Compare the results of these two requests:
const data1 = webix.ajax("data/nomencl/154"); //"{"id":154,"name":"John"}"
const data2 = webix.remote.nm.getNomenclature(154); //{id:154,name:"John"}After receiving the response of the AJAX request, you have to call JSON.parse(data1) to turn a string into an object. The response of the second request is already an object.
2. Requests with webix.remote are safer due to CSRF-security.
3. Several requests are sent as one, which makes operations faster.
For more details on services, read: