Java Platform, Enterprise Edition (Java EE) 8
The Java EE Tutorial

Previous Next Contents

The phonebilling Example Application

The phonebilling example application, located in the tut-install`/examples/batch/phonebilling/` directory, demonstrates how to use the batch framework in Java EE to implement a phone billing system. This example application processes a log file of phone calls and creates a bill for each customer.

The following topics are addressed here:

Architecture of the phonebilling Example Application

The phonebilling example application consists of the following elements.

  • A job definition file (phonebilling.xml) that uses the Job Specification Language (JSL) to define a batch job with two chunk steps. The first step reads call records from a log file and associates them with a bill. The second step computes the amount due and writes each bill to a text file.

  • A Java class (CallRecordLogCreator) that creates the log file for the batch job. This is an auxiliary component that does not demonstrate any key functionality in this example.

  • Two Java Persistence API (JPA) entities (CallRecord and PhoneBill) that represent call records and customer bills. The application uses a JPA entity manager to store instances of these entities in a database.

  • Three batch artifacts (CallRecordReader, CallRecordProcessor, and CallRecordWriter) that implement the first step of the application. This step reads call records from the log file, associates them with a bill, and stores them in a database.

  • Four batch artifacts (BillReader, BillProcessor, BillWriter, and BillPartitionMapper) that implement the second step of the application. This step is a partitioned step that gets each bill from the database, calculates the amount due, and writes it to a text file.

  • Two Facelets pages (index.xhtml and jobstarted.xhtml) that provide the front end of the batch application. The first page shows the log file that will be processed by the batch job, and the second page enables the user to check on the status of the job and shows the resulting bill for each customer.

  • A managed bean (JsfBean) that is accessed from the Facelets pages. The bean submits the job to the batch runtime, checks on the status of the job, and reads the text files for each bill.

The Job Definition File

The phonebilling.xml job definition file is located in the WEB-INF/classes/META-INF/batch-jobs/ directory. The file specifies three job-level properties and two steps:

<?xml version="1.0" encoding="UTF-8"?>
<job id="phonebilling" xmlns=""
        <property name="log_file_name" value="log1.txt"/>
        <property name="airtime_price" value="0.08"/>
        <property name="tax_rate" value="0.07"/>
    <step id="callrecords" next="bills"> ... </step>
    <step id="bills"> ... </step>

The first step is defined as follows:

<step id="callrecords" next="bills">
    <chunk checkpoint-policy="item" item-count="10">
        <reader ref="CallRecordReader"></reader>
        <processor ref="CallRecordProcessor"></processor>
        <writer ref="CallRecordWriter"></writer>

This step is a normal chunk step that specifies the batch artifacts that implement each phase of the step. The batch artifact names are not fully qualified class names, so the batch artifacts are CDI beans annotated with @Named.

The second step is defined as follows:

<step id="bills">
    <chunk checkpoint-policy="item" item-count="2">
        <reader ref="BillReader">
            <properties>                <property name="firstItem" value="#{partitionPlan['firstItem']}"/>                <property name="numItems" value="#{partitionPlan['numItems']}"/>            </properties>        </reader>
        <processor ref="BillProcessor"></processor>
        <writer ref="BillWriter"></writer>
        <mapper ref="BillPartitionMapper"/>
    <end on="COMPLETED"/>

This step is a partitioned chunk step. The partition plan is specified through the BillPartitionMapper artifact instead of using the plan element.

The CallRecord and PhoneBill Entities

The CallRecord entity is defined as follows:

public class CallRecord implements Serializable {
    @Id @GeneratedValue
    private Long id;
    private Date datetime;
    private String fromNumber;
    private String toNumber;
    private int minutes;
    private int seconds;
    private BigDecimal price;

    public CallRecord() { }

    public CallRecord(String datetime, String from,
            String to, int min, int sec)             throws ParseException { ... }

    public CallRecord(String jsonData) throws ParseException { ... }

    /* ... Getters and setters ... */

The id field is generated automatically by the JPA implementation to store and retrieve CallRecord objects to and from a database.

The second constructor creates a CallRecord object from an entry of JSON data in the log file using the JSON Processing API. Log entries look as follows:

{"datetime":"03/01/2013 04:03","from":"555-0101",

The PhoneBill entity is defined as follows:

public class PhoneBill implements Serializable {
    private String phoneNumber;
    @OneToMany(fetch = FetchType.EAGER, cascade = CascadeType.PERSIST)
    @OrderBy("datetime ASC")
    private List<CallRecord> calls;
    private BigDecimal amountBase;
    private BigDecimal taxRate;
    private BigDecimal tax;
    private BigDecimal amountTotal;

    public PhoneBill() { }

    public PhoneBill(String number) {
        this.phoneNumber = number;
        calls = new ArrayList<>();

    public void addCall(CallRecord call) {

    public void calculate(BigDecimal taxRate) { ... }

    /* ... Getters and setters ... *

The OneToMany annotation defines the relationship between a bill and its call records. The FetchType.EAGER attribute specifies that the collection should be retrieved eagerly. The CascadeType.PERSIST attribute indicates that the elements in the call list should be automatically persisted when the phone bill is persisted. The OrderBy annotation defines an order for retrieving the elements of the call list from the database.

The batch artifacts use instances of these two entities as items to read, process, and write.

For more information on the Java Persistence API, see Chapter 40, "Introduction to the Java Persistence API". For more information on the JSON Processing API, see Chapter 20, "JSON Processing".

The Call Records Chunk Step

The first step is composed of the CallRecordReader, CallRecordProcessor, and CallRecordWriter batch artifacts.

The CallRecordReader artifact reads call records from the log file:

public class CallRecordReader implements ItemReader {
    private ItemNumberCheckpoint checkpoint;
    private String fileName;
    private BufferedReader breader;
    JobContext jobCtx;

    /* ... Override the open, close, readItem,
     *     and checkpointInfo methods ... */

The open method reads the log_filename property and opens the log file with a buffered reader:

fileName = jobCtx.getProperties().getProperty("log_file_name");
breader = new BufferedReader(new FileReader(fileName));

If a checkpoint object is provided, the open method advances the reader up to the last checkpoint. Otherwise, this method creates a new checkpoint object. The checkpoint object keeps track of the line number from the last committed chunk.

The readItem method returns a new CallRecord object or null at the end of the log file:

public Object readItem() throws Exception {
    /* Read a line from the log file and
     * create a CallRecord from JSON */
    String callEntryJson = breader.readLine();
    if (callEntryJson != null) {
        return new CallRecord(callEntryJson);
    } else
        return null;

The CallRecordProcessor artifact obtains the airtime price from the job properties, calculates the price of each call, and returns the call object. This artifact overrides only the processItem method.

The CallRecordWriter artifact associates each call record with a bill and stores the bill in the database. This artifact overrides the open, close, writeItems, and checkpointInfo methods. The writeItems method looks like this:

public void writeItems(List<Object> callList) throws Exception {

    for (Object callObject : callList) {
        CallRecord call = (CallRecord) callObject;
        PhoneBill bill = em.find(PhoneBill.class, call.getFromNumber());
        if (bill == null) {
            /* No bill for this customer yet, create one */
            bill = new PhoneBill(call.getFromNumber());
        } else {
            /* Add call to existing bill */

The Phone Billing Chunk Step

The second step is composed of the BillReader, BillProcessor, BillWriter, and BillPartitionMapper batch artifacts. This step gets the phone bills from the database, computes the tax and total amount due, and writes each bill to a text file. Since the processing of each bill is independent of the others, this step can be partitioned and run in more than one thread.

The BillPartitionMapper artifact specifies the number of partitions and the parameters for each partition. In this example, the parameters represent the range of items each partition should process. The artifact obtains the number of bills in the database to calculate these ranges. It provides a partition plan object that overrides the getPartitions and getPartitionProperties methods of the PartitionPlan interface. The getPartitions method looks like this:

public Properties[] getPartitionProperties() {
    /* Assign an (approximately) equal number of elements
     * to each partition. */
    long totalItems = getBillCount();
    long partItems = (long) totalItems / getPartitions();
    long remItems = totalItems % getPartitions();

    /* Populate a Properties array. Each Properties element
     * in the array corresponds to a partition. */
    Properties[] props = new Properties[getPartitions()];

    for (int i = 0; i < getPartitions(); i++) {
        props[i] = new Properties();
                String.valueOf(i * partItems));
        /* Last partition gets the remainder elements */
        if (i == getPartitions() - 1) {
                    String.valueOf(partItems + remItems));
        } else {
    return props;

The BillReader artifact obtains the partition parameters as follows:

public class BillReader implements ItemReader {
    @Inject    @BatchProperty(name = "firstItem")    private String firstItemValue;    @Inject    @BatchProperty(name = "numItems")    private String numItemsValue;
    private ItemNumberCheckpoint checkpoint;    @PersistenceContext    private EntityManager em;    private Iterator iterator;

    public void open(Serializable ckpt) throws Exception {
        /* Get the range of items to work on in this partition */
        long firstItem0 = Long.parseLong(firstItemValue);
        long numItems0 = Long.parseLong(numItemsValue);

        if (ckpt == null) {
            /* Create a checkpoint object for this partition */
            checkpoint = new ItemNumberCheckpoint();
        } else {
            checkpoint = (ItemNumberCheckpoint) ckpt;

        /* Adjust range for this partition from the checkpoint */
        long firstItem = checkpoint.getItemNumber();
        long numItems = numItems0 - (firstItem - firstItem0);

This artifact also obtains an iterator to read items from the JPA entity manager:

/* Obtain an iterator for the bills in this partition */
String query = "SELECT b FROM PhoneBill b ORDER BY b.phoneNumber";
Query q = em.createQuery(query).setFirstResult((int) firstItem)
        .setMaxResults((int) numItems);
iterator = q.getResultList().iterator();

The BillProcessor artifact iterates over the list of call records in a bill and calculates the tax and total amount due for each bill.

The BillWriter artifact writes each bill to a plain text file.

The JavaServer Faces Pages

The index.xhtml page contains a text area that shows the log file of call records. The page provides a button for the user to submit the batch job and navigate to the next page:

    <h1>The Phone Billing Example Application</h1>
    <h2>Log file</h2>
    <p>The batch job analyzes the following log file:</p>
    <textarea cols="90" rows="25"
    <p> </p>
        <h:commandButton value="Start Batch Job"
                         action="#{jsfBean.startBatchJob()}" />

This page calls the methods of the managed bean to show the log file and submit the batch job.

The jobstarted.xhtml page provides a button to check the current status of the batch job and displays the bills when the job finishes:

<p>Current Status of the Job: <b>#{jsfBean.jobStatus}</b></p>
<h:dataTable var="_row" value="#{jsfBean.rowList}"
             border="1" rendered="#{jsfBean.completed}">
    <!-- ... show results from jsfBean.rowList ... -->
<!-- Render the check status button if the job has not finished -->
    <h:commandButton value="Check Status"
                     action="jobstarted" />

The Managed Bean

The JsfBean managed bean submits the job to the batch runtime, checks on the status of the job, and reads the text files for each bill.

The startBatchJob method of the bean submits the job to the batch runtime:

/* Submit the batch job to the batch runtime.
 * JSF Navigation method (return the name of the next page) */
public String startBatchJob() {
    jobOperator = BatchRuntime.getJobOperator();
    execID = jobOperator.start("phonebilling", null);
    return "jobstarted";

The getJobStatus method of the bean checks the status of the job:

/* Get the status of the job from the batch runtime */
public String getJobStatus() {
    return jobOperator.getJobExecution(execID).getBatchStatus().toString();

The getRowList method of the bean creates a list of bills to be displayed on the jobstarted.xhtml JSF page using a table.

Running the phonebilling Example Application

You can use either NetBeans IDE or Maven to build, package, deploy, and run the phonebilling example application.

The following topics are addressed here:

To Run the phonebilling Example Application Using NetBeans IDE

  1. Make sure that GlassFish Server has been started (see Starting and Stopping GlassFish Server).

  2. From the File menu, choose Open Project.

  3. In the Open Project dialog box, navigate to:

  4. Select the phonebilling folder.

  5. Click Open Project.

  6. In the Projects tab, right-click the phonebilling project and select Run.

    This command builds and packages the application into a WAR file, phonebilling.war, located in the target/ directory; deploys it to the server; and launches a web browser window at the following URL:


To Run the phonebilling Example Application Using Maven

  1. Make sure that GlassFish Server has been started (see Starting and Stopping GlassFish Server).

  2. In a terminal window, go to:

  3. Enter the following command to deploy the application:

    mvn install
  4. Open a web browser window at the following URL:


Previous Next Contents
Oracle Logo  Copyright © 2017, Oracle and/or its affiliates. All rights reserved.