StaxEventItemReader - skip XML fragments processed in previous Job executions - spring

Question

When working with CSV files and restarting a FAILED Job, a StepExecutionListner and associated beforeStep( ....) method can be used to position
the reader within the file. So the code could look something like:
public void beforeStep(StepExecution stepExecution) {
ExecutionContext executionContext = stepExecution.getExecutionContext();
if (executionContext.containsKey(getKey(LINES_READ_COUNT))) {
long lineCount = executionContext.getLong(getKey(LINES_READ_COUNT));
LineReader reader = getReader();
Object record = "";
while (reader.getPosition() < lineCount && record != null) {
record = readLine();
}
}
} // Or something similar
My question is how do you achieve the same thing when working with a StaxEventItemReader?
My batch_step_execution_context looks something like {"string":"StaxEventItemReader.read.count","int":6}. So in my case the
the first 5 XML fragment were successfully processed and upon restarting the Job I would like to start processing from XML fragment number 6 onwards.
Given the config below, how would I position the reader within the XML file?
<batch:job id="reportJob" restartable="true">
<batch:step id="step1">
<batch:tasklet>
<batch:chunk reader="xmlItemReader" writer="cvsFileItemWriter" processor="filterReportProcessor"
commit-interval="1">
</batch:chunk>
<batch:listeners>
<batch:listener ref="step1Listener" />
</batch:listeners>
</batch:tasklet>
</batch:step>
</batch:job>
<bean id="step1Listener" class="com.mkyong.listeners.Step1Listener" />
<bean id="filterReportProcessor" class="com.mkyong.processor.FilterReportProcessor" />
<bean id="xmlItemReader" class="org.springframework.batch.item.xml.StaxEventItemReader">
<property name="fragmentRootElementName" value="record" />
<property name="resource" value="classpath:xml/report.xml" />
<property name="unmarshaller" ref="reportUnmarshaller" />
</bean>
<!-- Read and map values to object, via jaxb2 -->
<bean id="reportUnmarshaller" class="org.springframework.oxm.jaxb.Jaxb2Marshaller">
<property name="classesToBeBound">
<list>
<value>com.mkyong.model.Report</value>
</list>
</property>
</bean>
Environment - spring-batch-core-2.2.0; spring-core-3.2.2
Test Input File
Convert a XML file into a CSV file.
<company>
<record refId="1001">
<name>mkyong</name>
<age>31</age>
<dob>31/8/1982</dob>
<income>200,000</income>
</record>
<record refId="1002">
<name>kkwong</name>
<age>30</age>
<dob>26/7/1983</dob>
<income>100,999</income>
</record>
<record refId="1003">
<name>joel</name>
<age>29</age>
<dob>21/8/1984</dob>
<income>1,000,000</income>
</record>
<record refId="1004">
<name>leeyy</name>
<age>29</age>
<dob>21/3/1984</dob>
<income>80,000.89</income>
</record>
<record refId="1005">
<name>Grant</name>
<age>29</age>
<dob>21/3/1984</dob>
<income>80,000.89</income>
</record>
</company>
Test Run 1
After processing two records in the input file, I forced a RunTimeException.
batch_job_execution --->> "FAILED";"FAILED";"java.lang.RuntimeException: Get me out of here!
batch_step_execution_context --->> {"string":"StaxEventItemReader.read.count","int":2}
Output CSV file --->> 1001,mkyong,31,31/08/1982,200000
1002,kkwong,30,26/07/1983,100999
Test Run 2
Process all "remaining data", so expecting .... refId="1003", refId="1004", refId="1005"
batch_job_execution --->> "COMPLETED";"COMPLETED";"''";"2015-01-25 16:18:08.587"
batch_step_execution_context --->> {"string":"StaxEventItemReader.read.count","int":6}
Output CSV file --->> 1001,mkyong,31,31/08/1982,200000
1002,kkwong,30,26/07/1983,100999
1003,joel,29,21/08/1984,1000000
1004,leeyy,29,21/03/1984,80000.89
1005,Grant,29,21/03/1984,80000.89
Test Result
Unfortunately it looks like the StaxEventItemReader is reading from the beginning of the file, rather than re-positioning itself based on the value of StaxEventItemReader.read.count which is set to 2 after the first test.

Answer 1

You don't need to configure anything, this is already the default behavior of the StaxEventItemReader. When it opens it repositions itself from the read count in the step execution context.