JdbcPagingItemReader混淆了SQL语句

Question

我正在使用JdbcPagingItemReader进行Spring Batch中的分区从属步骤。

以下是我的JobConfiguration for Partitioning步骤

@Bean
public Job loadConsumerJob() {
    return jobBuilderFactory.get(JOB_NAME)
            .incrementer(new RunIdIncrementer())
            .start(step1Master())
            .build();
}

public Step step1Master() {
    return stepBuilderFactory.get(STEP_ONE_MASTER_NAME)
            .<ConsumerNoMasterInterface, BillInterface> partitioner(STEP_ONE_NAME,jobTwoStepOneMasterConfiguration.jobTwoStepOnePartioner(null))
            .step(step1())
            .gridSize(jobTwoStepOneMasterConfiguration.jobTwoStepOneGridSize())
            .taskExecutor(jobTwoStepOneMasterConfiguration.jobTwoStepOneTaskExecutor())
            .listener(jobTwoStepOneMasterConfiguration.jobTwoStepOneMasterListener())
            .build();
}

public Step step1() {
    return stepBuilderFactory.get(STEP_ONE_NAME)
            .<ConsumerNoMasterInterface, BillInterface> chunk(CHUNK_SIZE)
            .reader(jobTwoStepOneConfiguration.jobTwoStepOneJdbcPagingItemReader(null,null,0,0,0))
            .processor(jobTwoStepOneConfiguration.jobTwoStepOneProcessor())
            .writer(jobTwoStepOneConfiguration.jobTwoStepOneWriter())
            .build();
}

这是我的QueryProvider

@Bean
public PostgresPagingQueryProvider jobTwoStepOneQueryProvider() {
    String methodName = "jobTwoStepOneQueryProvider() : ";
    logger.info(methodName + "called");

    PostgresPagingQueryProvider provider = new PostgresPagingQueryProvider();
    provider.setSelectClause("select * ");
    provider.setFromClause("from (\n" +
            "select cnm.*,row_number() over(order by id) as rownum from consumer_no_master cnm where \n" +
            "group_no=:groupNo AND consumer_no not in (select consumer_no from bill b where b.bill_month=:billMonth and deleted = false)\n" +
            ") as consumers ");
    provider.setWhereClause("where rownum between :fromRow AND :toRow ");

    Map<String,Order> sortKeys = new HashMap<>();
    sortKeys.put("rownum",Order.ASCENDING);
    provider.setSortKeys(sortKeys);

    return provider;
}

这是我的JdbcPagingItemReaderBuilder代码

@Bean
@StepScope
public JdbcPagingItemReader jobTwoStepOneJdbcPagingItemReader(@Value("#{jobParameters['GROUP_NO']}") String groupNo,
                                                                           @Value("#{jobParameters['BILL_MONTH']}") String billMonth,
                                                                           @Value("#{stepExecutionContext[fromRow]}") int fromRow,
                                                                           @Value("#{stepExecutionContext[toRow]}") int toRow,
                                                                           @Value("#{stepExecutionContext[limit]}") int limit) {
    String methodName = "jobTwoStepOneJdbcPagingItemReader() : ";
    logger.info(methodName + "called for groupNo " + groupNo + " billMonth " + billMonth + " fromRow " + fromRow + " toRow " + toRow + " with limit " + limit);

    Map<String, Object> parameterValues = new HashMap<>();
    parameterValues.put("groupNo", groupNo);
    parameterValues.put("billMonth",billMonth);
    parameterValues.put("fromRow",fromRow);
    parameterValues.put("toRow",toRow);

    return new JdbcPagingItemReaderBuilder<ConsumerNoMaster>()
            .name("consumerReader")
            .dataSource(dataSource)
            .queryProvider(jobTwoStepOneQueryProvider())
            .parameterValues(parameterValues)
            .rowMapper(new ConsumerNoMasterRowMapper())
            .pageSize(limit)
            .build();
}

这是我的分区功能

@Override
public Map<String, ExecutionContext> partition(int gridSize) {
    final String methodName = "partition() : ";
    logger.info(methodName + "called with gridSize " + gridSize + " chunk size " + chunkSize + " total size" + totalSize + " for group " + groupNo);

    Map<String, ExecutionContext> partitions = new HashMap<String, ExecutionContext>();

    if(gridSize <= 0 || totalSize <= 0 || chunkSize <= 0){
        logger.error(methodName + "Either of Param is <= 0");
        return partitions;
    }

    int work = new BigDecimal(totalSize).divide(new BigDecimal(gridSize),0,RoundingMode.CEILING).intValue();
    logger.info(methodName + "for group " + groupNo + " with total size" + totalSize + " and grid size " + gridSize + " unit of work for each cpu " + work);

    int fromRow = 1;
    int toRow = work;
    for(int i = 1; i <= gridSize; i++){
        logger.info(methodName + "For group " + groupNo + " with partition-" + i + " fromRow " + fromRow + " toRow " + toRow);
        ExecutionContext executionContext = new ExecutionContext();
        executionContext.put(JobTwoStepOneConfiguration.FROM_ROW,fromRow);
        executionContext.put(JobTwoStepOneConfiguration.TO_ROW,toRow);
        executionContext.put(JobTwoStepOneConfiguration.LIMIT,work);
        partitions.put("partition-" + i,executionContext);

        fromRow = toRow + 1;
        toRow = toRow + work;
    }

    return partitions;
}

当我使用1157的消费者大小运行我的工作时，JdbcPagingItemReader触发的SQL就是这个（从日志中捕获）

2018-06-15 13:56:41.728 DEBUG 14116 --- [SimpleAsyncTaskExecutor-4] o.s.b.i.database.JdbcPagingItemReader    : SQL used for reading remaining pages: [SELECT * FROM (
select cnm.*,row_number() over(order by id) as rownum from consumer_no_master cnm where 
group_no=:groupNo AND consumer_no not in (select consumer_no from bill b where b.bill_month=:billMonth and deleted = false)
) as consumers WHERE (rownum between :fromRow AND :toRow) AND ((rownum > :_rownum)) ORDER BY rownum ASC LIMIT 290]

2018-06-15 13:56:41.728 DEBUG 14116 --- [SimpleAsyncTaskExecutor-4] o.s.b.i.database.JdbcPagingItemReader    : Using parameterMap:{billMonth=MAY-2018, fromRow=1, toRow=290, groupNo=KSK13, _rownum=290}

我不知道为什么JdbcPagingItemReader在生成的SQL语句中设置_rownum。我没有将它设置为任何参数值映射。

我想要的是在传递fromRow和toRow值之间读取消费者为什么JdbcPagingItemReader在SQL中附加_rownum参数???

请指导我在这里做错了!!!