xuanlvhaoshao 2020-05-07
最近遇到一个需求,要按天统计一些数据,但之前系统已经运行大半年了,需要把之前的数据也做一个统计:
于是刚开始我是这样写的:
public void autoSaveData(RentalCostReportQueryVO queryVo) { Date startTime; Date endTime; if (queryVo.getStartTime() == null || queryVo.getEndTime() == null) { startTime = DateUtils.getCurrentDate(); endTime = DateUtils.getCurrentDate(); } else { startTime = queryVo.getStartTime(); endTime = queryVo.getEndTime().before(DateUtils.getCurrentDate()) ? queryVo.getEndTime() : DateUtils.getCurrentDate(); } Calendar startCalendar = Calendar.getInstance(); Calendar endCalendar = Calendar.getInstance(); startCalendar.setTime(startTime); endCalendar.setTime(endTime); List<RentalCostReport> rentalCostReportList = new ArrayList<>(); List<Date> dateList = new ArrayList<>(); do { Date currentTime = startCalendar.getTime(); //此处是循环执行的查询 List<RentalCostReport> rentalCostReportList1 = rentalCostMapper.queryRentalCostByDay(currentTime); //日期加1 startCalendar.add(Calendar.DAY_OF_MONTH, 1); //当前日期和结束日历日期比较,超过结束日期则终止 } while (!startCalendar.after(endCalendar)); List<RentalCostReport> rentalCostReportList2 = rentalCostMapper.queryFmAttendanceBlock(queryVo) .stream().filter(item -> null != item.getId()).collect(Collectors.toList()); List<RentalCostReport> rentalCostReportList3 = rentalCostMapper.queryStoreAndDeliverReport(queryVo); mergeRentalCostReportList(rentalCostReportList, rentalCostReportList2, rentalCostReportList3); if (CollectionUtils.isNotEmpty(rentalCostReportList)) { for (RentalCostReport rentalCostReport : rentalCostReportList) { rentalCostReport.setId(rentalCostReport.getId() + rentalCostReport.getLeaseCode()); } rentalCostMapper.batchSaveOrUpdate(rentalCostReportList); } }
其实很简单就是把某个时间段内遍历出具体是哪一天 ,然后把这个具体的天当做参数传递到mapper中,然后就循环查出每一天的集合 ,再把集合累加
do { Date currentTime = startCalendar.getTime(); //此处是循环执行的查询 List<RentalCostReport> rentalCostReportList1 = rentalCostMapper.queryRentalCostByDay(currentTime); //日期加1 startCalendar.add(Calendar.DAY_OF_MONTH, 1); //当前日期和结束日历日期比较,超过结束日期则终止 } while (!startCalendar.after(endCalendar));
这是一种做法,但这种做法很耗费时间,比如我要查询一年按天的统计就会循环执行365次,很耗费时间,于是我就在想,怎样能够批量传参(List<Date> 这个集合当中包含了具体的日期)然后批量返回结果
思路:在sql中 有个UNION 关键字 他是把相同结构的字段 合并返回
比如 SELECT #{currentDate} dateTime FROM A WHERE begin_time <= #{currentTime} AND end_time >= #{currentTime}
#{currentTime} 代表的是由mybatis传进来的参数 比如我传参进来是 ‘2020-05-04’ 但这个只能查询一天的
SELECT ‘2020-05-04’ dateTime FROM A WHERE begin_time <= ‘2020-05-04’ AND end_time >= ‘2020-05-04’
SELECT ‘2020-05-05’ dateTime FROM A WHERE begin_time <= ‘2020-05-05’ AND end_time >= ‘2020-05-05’
如上我如果要查询 ‘2020-05-05’ 就需要循环查询一次
但我用UNION 把他们连接起来呢?
SELECT ‘2020-05-04’ dateTime FROM A WHERE begin_time <= ‘2020-05-04’ AND end_time >= ‘2020-05-04’
UNION
SELECT ‘2020-05-05’ dateTime FROM A WHERE begin_time <= ‘2020-05-05’ AND end_time >= ‘2020-05-05’
这样就可以一次性查出来了
由此悟出来我可以使用Mybatis的 foreach 标签来循环一个集合 用UNION把他们联合起来,就形成了批量查询:
<select id="batchQueryRentalCostByDay" resultType="com.hope.saas.wms.entity.decision.report.RentalCostReport"> <foreach collection="list" item="currentTime" index="index" separator=" UNION "> SELECT CONCAT( a.tenant_id, b.temperature_layer_code, STR_TO_DATE( #{currentTime}, ‘%Y-%m-%d‘ )) id, STR_TO_DATE( #{currentTime}, ‘%Y-%m-%d‘ ) dateTime, b.temperature_layer_name temperatureLayerName, b.temperature_layer_code temperatureLayerCode, a.tenant_id tenantId, a.tenant_name tenantName FROM warehouse_lease_contract a LEFT JOIN leasing_information b ON a.id = b.warehouse_lease_contract_id WHERE begin_time <= #{currentTime} AND end_time >= #{currentTime} AND status_code = 1 GROUP BY lease_code, b.temperature_layer_code, a.tenant_id </foreach> </select>
运行起来效率相当快!