Spring Batch中BeanIO解析嵌套XML的类型转换错误解决

问题背景 #

基于Spring Boot的Spring Batch应用处理XML输入文件时，执行importUserJob的step1步骤抛出类型转换异常：com.test.springbatchapp.model.ContactList无法转换为com.test.springbatchapp.model.Contact。

XML输入结构包含外层query节点，内部results容器下有多个重复的contact节点，当前BeanIO配置将results映射为ContactList对象，但Step期望处理的是单个Contact对象，导致类型不匹配。

相关代码与配置 #

XML输入文件

<?xml version="1.0" encoding="UTF-8" ?>
<query>
    <id>123</id>
    <tracking>555</tracking>
    <results>
        <contact>
            <full_name>
                <first_name>John</first_name>
                <last_name>Doe</last_name>
            </full_name>
            <street>123 Main St</street>
            <city>Chicago</city>
            <state>IL</state>
            <zip>60610</zip>
        </contact>
        <contact>
            <full_name>
                <first_name>Jane</first_name>
                <last_name>Smith</last_name>
            </full_name>
            <street>123 Main St</street>
            <city>Miami</city>
            <state>FL</state>
        </contact>
    </results>
</query>

原BeanIO配置（beanio-configuration.xml）

<beanio xmlns="http://www.beanio.org/2012/03">
    <stream name="query" format="xml" strict="true" ignoreUnidentifiedRecords="true">
        <record name="id"></record>
        <record name="tracking"></record>
        <record name="results" class="com.test.springbatchapp.model.ContactList">
            <segment name="contact" collection="list" class="com.test.springbatchapp.model.Contact" occurs="0+">
                <segment name="full_name">
                    <field name="firstName" xmlName="first_name" maxLength="20" />
                    <field name="lastName" xmlName="last_name" maxLength="30" />
                </segment>
                <field name="street" maxLength="30" />
                <field name="city" maxLength="25" />
                <field name="state" minLength="2" maxLength="2" />
                <field name="zip" regex="\d{5}" minOccurs="0" default="" />
            </segment>
        </record>
    </stream>
</beanio>

实体类

Contact.java：

@Data
@AllArgsConstructor
@NoArgsConstructor
public class Contact {
    
    private String firstName;
    private String lastName;
    private String street;
    private String city;
    private String state;
    private String zip;
}

ContactList.java：

@Data
@AllArgsConstructor
@NoArgsConstructor
public class ContactList {
    
    private List<Contact> contact;
}

原Reader配置

@Bean
public ItemReader<Contact> reader() {
    BeanIOFlatFileItemReader<Contact> reader = new BeanIOFlatFileItemReader<>();
    try {
        reader.setResource(new FileSystemResource(fileInputContact));
        reader.setStreamName(inputContactStreamName);
        reader.setStreamMapping(new ClassPathResource(beanIoConfigurationXmlPath));
        reader.setStreamFactory(StreamFactory.newInstance());
        reader.getLineNumber();
        reader.afterPropertiesSet();
    } catch (Exception e) {
        log.error("ERROR: An issue occurred in the BeanIO Item Reader:: {} {}", e.getMessage(), e.getStackTrace());
    }
    
    return reader;      
}

解决方案 #

方案一：修改BeanIO配置，直接返回单个Contact对象（推荐） #

调整BeanIO配置，忽略外层无关节点，直接将每个contact节点映射为Contact对象，让Reader返回Step期望的类型：

修改后的beanio-configuration.xml：

<beanio xmlns="http://www.beanio.org/2012/03">
    <stream name="query" format="xml" strict="true" ignoreUnidentifiedRecords="true">
        <!-- 忽略外层query、id、tracking节点，直接读取results下的contact -->
        <record name="query" ignore="true">
            <field name="id" ignore="true"/>
            <field name="tracking" ignore="true"/>
            <segment name="results" ignore="true">
                <record name="contact" class="com.test.springbatchapp.model.Contact" occurs="0+">
                    <segment name="full_name">
                        <field name="firstName" xmlName="first_name" maxLength="20" />
                        <field name="lastName" xmlName="last_name" maxLength="30" />
                    </segment>
                    <field name="street" maxLength="30" />
                    <field name="city" maxLength="25" />
                    <field name="state" minLength="2" maxLength="2" />
                    <field name="zip" regex="\d{5}" minOccurs="0" default="" />
                </record>
            </segment>
        </record>
    </stream>
</beanio>

说明：此配置会跳过query、id、tracking、results这些外层节点，直接识别并映射每个contact节点为Contact对象，Reader返回的类型与Step期望的Contact完全匹配，无需修改其他配置。

方案二：保留ContactList结构，通过Processor拆分列表 #

如果需要保留ContactList的映射结构，可调整Reader泛型，并通过ItemProcessor将ContactList拆分为单个Contact对象：

1. 修改Reader泛型为ContactList：

@Bean
public ItemReader<ContactList> reader() {
    BeanIOFlatFileItemReader<ContactList> reader = new BeanIOFlatFileItemReader<>();
    try {
        reader.setResource(new FileSystemResource(fileInputContact));
        reader.setStreamName(inputContactStreamName);
        reader.setStreamMapping(new ClassPathResource(beanIoConfigurationXmlPath));
        reader.setStreamFactory(StreamFactory.newInstance());
        reader.getLineNumber();
        reader.afterPropertiesSet();
    } catch (Exception e) {
        log.error("ERROR: An issue occurred in the BeanIO Item Reader:: {} {}", e.getMessage(), e.getStackTrace());
    }
    
    return reader;      
}

2. 修改Step配置与添加拆分Processor：

@Bean
public Step step1(FlatFileItemWriter<Contact> writer) {
    return stepBuilderFactory.get("step1")
      .<ContactList, Contact> chunk(10)
      .reader(reader())
      .processor(contactListSplitter())
      .writer(writer)
      .build();
}

@Bean
public ItemProcessor<ContactList, List<Contact>> contactListSplitter() {
    // 返回ContactList中的contact列表，Spring Batch会自动展开列表元素到Chunk中
    return contactList -> contactList.getContact();
}

说明：此方案下，Reader返回ContactList对象，Processor将其拆分为Contact列表，Spring Batch会自动将列表中的每个Contact作为独立元素传入Writer处理。

内容的提问来源于stack exchange，提问作者Somebody