项目背景
问题：
由上图可以知道，“页面需要根据试卷id获取全部的试卷信息”，由下面代码可以看出根据返回的“returnPaperDetail”进行开始进行嵌套循环操作。
图解：
具体代码如下
/*** 功能：queryAllQuestion* 描述：加载页面查询出所有表的数据* 主要实体：ExaminationPaper（4张表的联合：questionMain，questionSub,paperMain,paperSub）* 作者:十一期 谭倩倩* 时间：2016-6-18* 修改人：* 修改时间：* @param request* @param response*/@RequestMapping("/queryAllQuestion")public void queryAllQuestion(HttpServletRequest request,HttpServletResponse response){//1.抽取paperMainExaminationPaper examinationPaper=new ExaminationPaper();String paperMainId=paperId.replace("'", "");//暂时使用假数据String dataBaseName="itoo_exam";PaperMain RepaperMain=paperMainBean.queryPaperMainById(paperMainId, dataBaseName);examinationPaper.setId(paperMainId);examinationPaper.setPaperName(RepaperMain.getComment());examinationPaper.setPaperScore(RepaperMain.getScore());//2. 抽取paperdetail和questionMain数据（嵌套循环）List<PaperDetail> returnPaperDetail=this.ReturnqueryPaperDetail(paperMainId, dataBaseName);List<QuestionMain> RequestionMainList=new ArrayList<QuestionMain>();List<QuestionSub> RequestionSubList=new ArrayList<QuestionSub>();//循环paperdetailfor (PaperDetail returnSinglePaperDetail : returnPaperDetail){RequestionMainList=this.ReturnQuestionMain(returnSinglePaperDetail.getId(), dataBaseName);RequestionMainList = orderByQuestionMainList(RequestionMainList);for (QuestionMain returnSinglequestionMain : RequestionMainList){RequestionSubList=this.ReturnQuestionSub(returnSinglequestionMain.getId(), dataBaseName);returnSinglequestionMain.setOptions(RequestionSubList);}//查询出来的组件List<QuestionTypeDetail> lstQuestionTypeDetail = paperMainBean.queryComponentById(returnSinglePaperDetail.getQuestionTypeId(), dataBaseName);returnSinglePaperDetail.setLstQuestionTypeDetail(lstQuestionTypeDetail);//把组件的英文名称装到“returnSinglePaperDetail”里面returnSinglePaperDetail.setLstQuestionMain(RequestionMainList);}examinationPaper.setPaperDetails(returnPaperDetail);jacksonJson.beanToJson(response, examinationPaper);}/*** @param RequestionMainList* @return*/private List<QuestionMain> orderByQuestionMainList(List<QuestionMain> RequestionMainList) {/**对大小题进行排序思路：// * ①定义一个 List<QuestionMain> bigLittleQuestion 用于存放排好序的大小题。// * ②判断查出来的题型是大小题，则先筛选大题。// * ③如果是大题，将其放到bigLittleQuestion中。// * 然后遍历所有题目，找到该大题对应的所有小题。放到bigLittleQuestion中。// *// *// */………………………………………………………………}
【解决方案】
一套试卷有多个题型，那么我就设置每个题型为一个线程。
图示：
代码如下：
大体是实例化一个线程池，根据题目数量的多少来添加多少个线程，一个题型代表一个线程，最后遍历线程结果。
public void queryAllQuestion(HttpServletRequest request,HttpServletResponse response){//1.抽取paperMainExaminationPaper examinationPaper=new ExaminationPaper();String paperMainId=paperId.replace("'", "");//暂时使用假数据String dataBaseName="itoo_exam";PaperMain RepaperMain=paperMainBean.queryPaperMainById(paperMainId, dataBaseName);examinationPaper.setId(paperMainId);examinationPaper.setPaperName(RepaperMain.getComment());examinationPaper.setPaperScore(RepaperMain.getScore());//2. 抽取paperdetail和questionMain数据（嵌套循环）List<PaperDetail> returnPaperDetail=this.ReturnqueryPaperDetail(paperMainId, dataBaseName);//测试加上线程的时间long startTime=System.currentTimeMillis();int count=returnPaperDetail.size();//创建一个线程池ExecutorService executorService = Executors.newCachedThreadPool();List<Future<PaperDetail>> resultList = new ArrayList<Future<PaperDetail>>();for (PaperDetail returnSinglePaperDetail : returnPaperDetail) {//使用ExecutorService执行Callable类型的任务，并将结果保存在future变量中Future<PaperDetail> future = executorService.submit(new myCallable(returnSinglePaperDetail));//将任务执行结果存储到List中resultList.add(future);}List<PaperDetail> listPapaerDeatil=new ArrayList<PaperDetail>();//遍历任务的结果for (Future<PaperDetail> fp: resultList) {try {PaperDetail enPaperDetail =fp.get();listPapaerDeatil.add(enPaperDetail);} catch (InterruptedException | ExecutionException e) {// TODO Auto-generated catch blocke.printStackTrace();}//启动一次顺序关闭，执行以前提交的任务，但不接受新任务。如果已经关闭，则调用没有其他作用。executorService.shutdown();}examinationPaper.setPaperDetails(listPapaerDeatil);long endTime=System.currentTimeMillis();System.out.println(endTime-startTime);jacksonJson.beanToJson(response, examinationPaper);}
/*** 多线程进行题干和选项查询。* @author 十一期 谭倩倩**/@SuppressWarnings("unchecked")public class myCallable implements Callable<PaperDetail>{private PaperDetail paperDetail;public myCallable(PaperDetail paperDetail) {this.paperDetail = paperDetail;}String dataBaseName="itoo_exam";public PaperDetail call(){List<QuestionMain> RequestionMainList=questionMainBean.queryQuestionbankList(paperDetail.getId(), dataBaseName);RequestionMainList = orderByQuestionMainList(RequestionMainList);for (QuestionMain returnSinglequestionMain : RequestionMainList){List<QuestionSub> RequestionSubList=questionSubBean.queryQuestionbankOptionsList(returnSinglequestionMain.getId(), dataBaseName);returnSinglequestionMain.setOptions(RequestionSubList);}//查询出来的组件List<QuestionTypeDetail> lstQuestionTypeDetail = paperMainBean.queryComponentById(paperDetail.getQuestionTypeId(), dataBaseName);paperDetail.setLstQuestionTypeDetail(lstQuestionTypeDetail);//把组件的英文名称装到“returnSinglePaperDetail”里面paperDetail.setLstQuestionMain(RequestionMainList);return paperDetail;}}
【结果】
没有用线程的结果是：
（第一套测试卷子）
（第二套测试卷子）
用了线程的结果是：
（第一套测试卷子）
（第二套测试卷子）
【总结】
还需要优化的性能：
1. 算法优化
（需要优化的代码）
private List<QuestionMain> orderByQuestionMainList(List<QuestionMain> RequestionMainList) {/**对大小题进行排序思路：// * ①定义一个 List<QuestionMain> bigLittleQuestion 用于存放排好序的大小题。// * ②判断查出来的题型是大小题，则先筛选大题。// * ③如果是大题，将其放到bigLittleQuestion中。// * 然后遍历所有题目，找到该大题对应的所有小题。放到bigLittleQuestion中。// *// * 仍需要优化// */if (RequestionMainList!=null && RequestionMainList.size()>1 ){if ( RequestionMainList.get(0).getIsParentQuestion()==1 || RequestionMainList.get(0).getParentQuestionId()!=null){//筛选是大小题List<QuestionMain> bigLittleQuestion=new ArrayList<QuestionMain>();for (int i = 0; i < RequestionMainList.size(); i++){if (RequestionMainList.get(i).getIsParentQuestion()==1){//查找出所有的大题干bigLittleQuestion.add(RequestionMainList.get(i));//查找该大题干下面所有的小题干for (int j = 0; j < RequestionMainList.size(); j++){//匹配该大题干下面的小题干if(RequestionMainList.get(j).getParentQuestionId()!=null){if (RequestionMainList.get(j).getParentQuestionId().equals(RequestionMainList.get(i).getId())){bigLittleQuestion.add(RequestionMainList.get(j));}}}}}RequestionMainList=bigLittleQuestion;}}return RequestionMainList;}
2. 批量查询或者缓存
（需要优化的代码）
for (QuestionMain returnSinglequestionMain : RequestionMainList){List<QuestionSub> RequestionSubList=questionSubBean.queryQuestionbankOptionsList(returnSinglequestionMain.getId(), dataBaseName);