Sorting huge data using ZK

From Documentation
DocumentationSmall Talks2011MarchSorting huge data using ZK
Sorting huge data using ZK

Author
Jumper Chen, Engineer, Potix Corporation
Date
March 10, 2011
Version
ZK 5.0.6

WarningTriangle-32x32.png This page is under construction, so we cannot guarantee the accuracy of the content!

Introduction

In the previous article, we taught you how to implement a paging with huge data. Here, we will extend that example to simplify its model implementation and add a sorting function to work with the huge data. In the following example, the paging navigation and the sorting function are handled by Database query result so that we can reduce the memory usage and load the data on demand.

Demo

Creating the database

If you have already seen the previous article, you can jump to the #SortingPagingListModel section directly.

We create a database named livedata and create a table named tblusers. We will provide the following columns, id (INT, AI, PK), first_name (VARCHAR), last_name(VARCHAR) and phone(VARCHAR).

The SQL statement to create the table is provided below


CREATE TABLE tblusers (id int(10) NOT NULL AUTO_INCREMENT, 
                                      first_name varchar(200) NOT NULL, 
                                      last_name varchar(200) NOT NULL, 
                                      phone varchar(20) NOT NULL, PRIMARY KEY (id) ) 
                                      ENGINE=InnoDB DEFAULT CHARSET=utf8;

The database screen

We have introduced a database screen which allows you to quickly setup and test your database connection. This is the first page that you arrive at when running the sample, if you have any problems you can manually edit the DatabaseInteraction class and recompile it.

Interacting with the database

Linking the application with the database is the easy part and involves creating a DAO to handle the interaction. The DAO, in this instance, only consists of functions to insert and retrieve news items from the database. Please refer to the UserDAO for more details.

The connection is handled via the MySQL J/Connector with a new connection being made for each operation. Please note that a new connection for each operation is NOT the best method of implementation, a connection pool would be far more efficient so please do not use the code for production level software. Building the model

The model needs to represent our user utilizing Java’s Object Orientated features, hence we implement a Java class named User which includes getters and setters for every field present in the database.


The code within the User class is included below:

public class User {
 
          private int _id;
          private String _firstName;
          private String _lastName;
          private String _phone;
          
          public String getFirstName() {
                   return _firstName;
          }
          public void setFirstName(String firstName) {
                   this._firstName = firstName;
          }
          public String getLastName() {
                   return _lastName;
          }
          public void setLastName(String lastName) {
                   this._lastName = lastName;
          }
          public String getPhone() {
                   return _phone;
          }
          public void setPhone(String phone) {
                   this._phone = phone;
          }
          public void setId(int id) {
                   this._id = id;
          }
          public int getId() {
                   return _id;
          }
}

With the User class implemented along with the functions to retrieve them from the database we need to construct the model.


SortingPagingListModel

Building the UI

The basic GUI is very simple and consists of two parts

  1. The button which generates 1,000,000 database entries
  2. The paging area (including Grid and paging components)


Sorting Column Fields within Grid

One of the most important features of this small talk is that we use Grid in a paging mold and specify the sort attribute with "auto(field_name)".

The markup below is the ZUL code from our hugedata2.zul

<window title="Huge data demo" width="720px" border="normal" apply="org.zkoss.zklargelivelist.controllers.MainUserController2">
          <label value="If you need to generate 1,000,000 entries please press this button!" />
          <button id="btnGenerate" label="Generate database entries" />
          <separator />
          <label value="Press this to delete the page" />
          <button id="btnDeletepage" label="Delete the page" />
          <separator />
          
          <grid id="dataGrid" width="700px" mold="paging" pagingPosition="top">
                   <columns>
                             <column label="id" sort="auto(id)" width="70px"/>
                             <column label="firstName" sort="auto(first_name)"/>
                             <column label="lastName" sort="auto(last_name)"/>
                             <column label="phone" sort="auto(phone)"/>
                   </columns>
          </grid>
</window>


Controller functionality

The controller MainUserController2.java handles one GUI events and overrides the doAfterCompose method. Firstly let us deal with the doAfterCompose method and explain what it does. The method is called after the components have been initialized. This allows us to manipulate the components before the user interacts with them.


doAfterCompose

          public void doAfterCompose(Component comp) throws Exception {
                   super.doAfterCompose(comp);
 
                   dataGrid.setRowRenderer(new RowRenderer(){
                             @Override
                             public void render(Row row, Object data) throws Exception {
                                      User usr = (User) data;
                                      row.appendChild(new Label(String.valueOf(usr.getId())));
                                      row.appendChild(new Label(usr.getFirstName()));
                                      row.appendChild(new Label(usr.getLastName()));
                                      row.appendChild(new Label(usr.getPhone()));
                             }});
                   dataGrid.setModel(new SortingPagingListModel());
          }


We create our own custom row renderer to format the data and then create a SortingPagingListModel for it. Notice the speed and low memory usage of this as we only pull the amount of users that we require for one page and hold that in memory!

GUI events

          public void onClick$btnGenerate() {      
                   int numberOfUsers = DatabaseInformation.dao.getUserCount();
                   final int numberOfEntriesRequired = 1000000;
                   
                   int usersRequired = numberOfEntriesRequired - numberOfUsers;
                   
                   if(usersRequired > 0) {
                             User[] users = new User[usersRequired];
                                                         
                             for(int i=0; i<usersRequired; i++) {
                                      users[i] = GenerateData.generateUser();
                             }
                                      
                             DatabaseInformation.dao.insertUser(users);
                                                         
                             refreshModel(userPaging.getActivePage());
                   }
                   else
                   {
                             try {
                                      Messagebox.show("There are already 1,000,000 entries in the database!");
                             } catch (InterruptedException e) {
                                      e.printStackTrace();
                             }
                   }
          }


The other two functions are related to handling GUI events, the first is onClick$btnGenerate which is used to generate 1,000,000 entries in our database.

Configuration

Turn on the setting of Render-on-Demand (ROD) for this example. You can refer to this setting.

Download



Copyright © Potix Corporation. This article is licensed under GNU Free Documentation License.