querydsl+jpa编程中如何使用函数

原生的sql中使用的函数，具体的sql实现如下：

select tt.extend_id,
       tt.extend_name,
       wm_concat(distinct tt.extend_code),
       sum(rr.annuali_count),
       count(distinct rr.staff_id)
  from tm_duct_whole rr
  left join tm_duct_extend_data tt
    on rr.industry_type = tt.extend_code
 where tt.extend_type = 3
   and rr.is_close = 0
   and rr.is_login = 1
   and rr.industry_type is not null
 group by tt.extend_id, tt.extend_name

转化为jpa实现的编程，借助模板使用函数。具体的操作如下所示：

//编译生成的实体类
QTmDuctWhole tmDuctWhole = QTmDuctWhole.tmDuctWhole;
QBasicData basicData = QBasicData.basicData;
//求和
SimpleTemplate<Double> annualiCount = Expressions
    .template(Double.class, "round(sum(ANNUALI_COUNT))");
//进行字符串拼接    
SimpleTemplate<String> extendCode = Expressions
    .template(String.class, "to_char(wm_concat(distinct EXTEND_CODE))");
//计算总人数    
SimpleTemplate<Integer> userCount = Expressions
    .template(Integer.class, " count(distinct staff_id)");
List<TrackBean> industryList = secondaryJpaSqlQuery()
    .select(Projections.bean(TrackBean.class,
        basicData.extendId.as("extendId"),
        basicData.extendName.as("extendName"),
        extendCode.as("extendCode"),
        annualiCount.as("annualiCount"),
        userCount.as("count")))
    .from(tmDuctWhole).leftJoin(basicData)
    .on(tmDuctWhole.industryType.eq(basicData.extendCode))
    .where(basicData.extendType.eq(3)
        .and(tmDuctWhole.isClose.eq(0))
        .and(tmDuctWhole.isLogin.eq(1))
        .and(tmDuctWhole.industryType.isNotEmpty()))
    .groupBy(basicData.extendId, basicData.extendName)
    .fetch();