Mybatis流式查询并实现将结果分批写入文件

    /**
     * 流式查询,全量导出
     *
     * @param req  查询条件
     * @param size 单个文件数据最大条数
     * @return
     */
    @ApiOperation(value = "流式查询,全量导出")
    @GetMapping("/streamAll")
    public BaseResultModel streamAll(ReqBillRecordBackQuery req, Integer size) {
        try {
            billRecordBackService.streamAll(req, size);
        } catch (Exception e) {
            throw new RuntimeException(e);
        }
        return BaseResultModel.success();
    }

 @Override
    @Transactional
    public void streamAll(ReqBillRecordBackQuery req, Integer size) throws Exception {
        exportXml(req,size);
    }
    private void exportXml(ReqBillRecordBackQuery req, Integer size) throws Exception{
        //文件内容行数
        Integer in = 0;
        //文件名称
        Integer fileName=0;
        String name = "exportTest";
        String suf =".txt";
        String path = "H:\\新建文件夹\\新建文件夹\\export\\";
        File ff = new File(path);
        //递归删除目录中的所有文件和子目录，而不删除目录本身。
        FileUtils.cleanDirectory(ff);
        File fe = new File(path+name+fileName+suf);
        //删除此抽象路径名表示的文件或目录
        //mkdirs()可以建立多级文件夹， mkdir()只会建立一级的文件夹
//        fe.mkdirs();
        //获取文件输出列
        BufferedWriter bufferedWriter=new BufferedWriter(new FileWriter(fe));
        StringBuilder sb = new StringBuilder();
        Cursor<BillRecordBack> billRecordBacks = mapper.streamAll(req);
        for (BillRecordBack bil : billRecordBacks) {
            sb.append(bil).append("\n");
            in++;
            if (in>=size){
                in=0;
                fileName++;
                fe = new File(path+name+fileName+suf);
                bufferedWriter = new BufferedWriter(new FileWriter(fe));
            }
            bufferedWriter.write(sb.toString());
            //将StringBuilder数据重置
            sb.setLength(0);
        }
        //最后需要自己关闭流
        billRecordBacks.close();
        bufferedWriter.close();
    }

    <select id="streamAll" resultType="com.psh.hik.entity.BillRecordBack" fetchSize="5000">
        select t_id,r_id,r_time,r_number,descd,deleted,ctime,crname,mtime,chname from  bill_record_back
        <where>
            <if test="null != param.rTime and ''!= param.rTime">
              ctime = #{param.rTime}
            </if>
            <if test="null != param.rNumber and ''!= param.rNumber">
                ctime = #{param.rNumber}
            </if>
        </where>
    </select>

    private void exportNote(ReqBillRecordBackQuery req, Integer size) throws Exception{
        //lambda表达式访问外部变量有一个非常重要的限制：变量不可变（只是引用不可变，而不是真正的不可变）,AtomicInteger是一个提供原子操作的Integer类，通过线程安全的方式操作加减。
        //文件内容行数
        AtomicInteger in = new AtomicInteger(1);
        //文件名称
        AtomicInteger fileName= new AtomicInteger(0);
        String name = "exportTest";
        String suf =".txt";
        String path = "H:\\新建文件夹\\新建文件夹\\export\\";
        File ff = new File(path);
        //递归删除目录中的所有文件和子目录，而不删除目录本身。
        FileUtils.cleanDirectory(ff);
        AtomicReference<File> fe = new AtomicReference<>(new File(path + name + fileName + suf));
        AtomicReference<BufferedWriter> bufferedWriter= new AtomicReference<>(new BufferedWriter(new FileWriter(fe.get())));
        StringBuilder sb = new StringBuilder();
        mapper.exportNote(req,resultContext -> {
            try {
                if (fileName.get()>=20){
                    return;
                }
                BillRecordBack resultObject = resultContext.getResultObject();
                sb.append(resultObject).append("\n");
                //a.incrementAndGet(); 先+1，再返回,a.getAndIncrement()先返回，再 +1
                in.getAndIncrement();
                System.out.println(in);
                if (in.get() >=size){
                    in.set(0);
                    fileName.getAndIncrement();
                    fe.set(new File(path + name + fileName + suf));
                    bufferedWriter.set(new BufferedWriter(new FileWriter(fe.get())));
                }
                bufferedWriter.get().write(sb.toString());
                //将StringBuilder数据重置
                sb.setLength(0);
            }catch (Exception e){
                throw new RuntimeException(e);
            }
        });
        bufferedWriter.get().close();
    }

    @Select("select t_id,r_id,r_time,r_number,descd,deleted,ctime,crname,mtime,chname from  bill_record_back")
    //这个注解是设定每次流式查询的iterator大小的，这里是1000条 ,ResultSetType.FORWARD_ONLY 只允许游标向下移动
    @Options(resultSetType = ResultSetType.FORWARD_ONLY, fetchSize = 5000)
    @ResultType(BillRecordBack.class)
    void exportNote(ReqBillRecordBackQuery req, ResultHandler<BillRecordBack> handler);

<dependency>
  <groupId>org.mybatis.spring.boot</groupId>
  <artifactId>mybatis-spring-boot-starter</artifactId>
  <version>1.1.1</version>
</dependency>

<select id="selectGxids" resultType="java.lang.String" fetchSize="1000">
   SELECT gxid from t_gxid
 </select>

package flowselect;
import org.apache.ibatis.session.ResultContext;
import org.apache.ibatis.session.ResultHandler;
import java.util.Set;
public class GxidResultHandler implements ResultHandler<String> {
  // 这是每批处理的大小
  private final static int BATCH_SIZE = 1000;
  private int size;
  // 存储每批数据的临时容器
  private Set<String> gxids;
  public void handleResult(ResultContext<? extends String> resultContext) {
    // 这里获取流式查询每次返回的单条结果
    String gxid = resultContext.getResultObject();
    // 你可以看自己的项目需要分批进行处理或者单个处理，这里以分批处理为例
    gxids.add(gxid);
    size++;
    if (size == BATCH_SIZE) {
      handle();
    }
  }
  private void handle() {
    try {
      // 在这里可以对你获取到的批量结果数据进行需要的业务处理
    } finally {
      // 处理完每批数据后后将临时清空
      size = 0;
      gxids.clear();
    }
  }
  // 这个方法给外面调用，用来完成最后一批数据处理
  public void end(){
    handle();// 处理最后一批不到BATCH_SIZE的数据
  }
}

package flowselect;
import org.mybatis.spring.SqlSessionTemplate;
import org.springframework.beans.factory.annotation.Autowired;
@Service
public class ServiceImpl implements Service {
  @Autowired
  SqlSessionTemplate sqlSessionTemplate;
  public void method(){
    GxidResultHandler gxidResultHandler = new GxidResultHandler();
    sqlSessionTemplate.select("flowselect.Mapper.selectGxids", gxidResultHandler);
    gxidResultHandler.end();
  }
}