这样的语句如何优化Temp空间 现在它要用30G哦

本文通过具体的SQL查询案例,展示了如何通过合理的表连接顺序和分组策略来减少临时表空间的使用,进而提高查询效率。文中涉及的技术包括SQL优化技巧、表连接和分组操作等。
select  /*+PARALLET(T_BASE_USER_BUY_INFO,4)*/ a.f_yearweek,a.f_lotbigclass,b.ExpeNum,a.salay,b.countrysalay,a.bonus,b.countrybonus
  from
      (
          select TO_CHAR(b.f_DATE,'YYYYMMWW') as f_yearweek,c.f_lotbigclass,F_FlatName,
          sum(F_BuyMoney) Salay,
          sum(F_TaxBefforBonus) Bonus
          from T_Base_User_Buy_Info b
          left join T_Base_Lot_Send_Bonus s
             on b.F_LotID=s.f_lotid and s.f_wtype=b.f_wtype  and s.f_expect=b.f_expect
          inner join T_base_Flat_Type t
             on t.f_buymodename=b.F_BuyChannel and t.F_DEPARTMENT=1 and f_flatname='主站'
          inner join ( select distinct f_lotbigclass,f_lottype from t_base_lotclass t where f_classsort=1) C
             on b.F_LotID=c.f_lottype
          where b.f_date between to_date('2007-01-01 00:00:00','yyyy-mm-dd hh24:mi:ss') and to_date('2007-01-10 00:00:00','yyyy-mm-dd hh24:mi:ss')
          and s.f_date between to_date('2007-01-01 00:00:00','yyyy-mm-dd hh24:mi:ss') and to_date('2007-01-10 00:00:00','yyyy-mm-dd hh24:mi:ss')         
          group by TO_CHAR(b.f_DATE,'YYYYMMWW'),c.f_lotbigclass,F_FlatName
      ) a inner join
      (
         select TO_CHAR(b.f_DATE,'YYYYMMWW') as f_yearweek,c.f_lotbigclass,
          count(distinct b.F_Expect) ExpeNum,
          sum(b.F_BuyMoney) countrySalay,
          sum(s.F_TaxBefforBonus) countryBonus
          from T_Base_User_Buy_Info b
          left join T_Base_Lot_Send_Bonus s
             on b.F_LotID=s.f_lotid and s.f_wtype=b.f_wtype  and s.f_expect=b.f_expect
          inner join T_base_Flat_Type t
             on t.f_buymodename=b.F_BuyChannel and t.F_DEPARTMENT=1
           inner join ( select distinct f_lotbigclass,f_lottype from t_base_lotclass t where f_classsort=1) C
             on b.F_LotID=c.f_lottype
          where b.f_date between to_date('2007-01-01 00:00:00','yyyy-mm-dd hh24:mi:ss') and to_date('2007-01-10 00:00:00','yyyy-mm-dd hh24:mi:ss')
          and s.f_date between to_date('2007-01-01 00:00:00','yyyy-mm-dd hh24:mi:ss') and to_date('2007-01-10 00:00:00','yyyy-mm-dd hh24:mi:ss')         
          group by TO_CHAR(b.f_DATE,'YYYYMMWW'),c.f_lotbigclass,F_FlatName
      ) b on a.f_yearweek=b.f_yearweek and a.f_lotbigclass=b.f_lotbigclass ;
     
执行计划

 
最后把两个大表现统计后在LEFT JOIN  那样占临时表空间就少了
  select a.f_yearweek,c.f_lotbigclass,sum(ExpeNum) ExpeNum ,sum(countrySalay) countrySalay,sum(countryBonus) countryBonus
          from
          (
              select TO_CHAR(f_DATE,'YYYYMMWW') as f_yearweek ,F_LotID, count(distinct F_Expect) ExpeNum, sum(F_BuyMoney) countrySalay
              from T_Base_User_Buy_Info
              where f_date between to_date('2007-01-01 00:00:00','yyyy-mm-dd hh24:mi:ss') and to_date('2007-01-10 00:00:00','yyyy-mm-dd hh24:mi:ss')
              group by TO_CHAR(f_DATE,'YYYYMMWW'),F_LotID
          ) a
          left join
          (
              select TO_CHAR(f_date,'YYYYMMWW') as f_yearweek ,F_LotID,sum(F_TaxBefforBonus) countryBonus
              from T_Base_Lot_Send_Bonus
              where f_date between to_date('2007-01-01 00:00:00','yyyy-mm-dd hh24:mi:ss') and to_date('2007-01-10 00:00:00','yyyy-mm-dd hh24:mi:ss')         
              group by TO_CHAR(f_date,'YYYYMMWW'),F_LotID
          )b  on a.f_yearweek=b.f_yearweek and a.f_lotid=b.f_lotid
          inner join
          (
           select distinct f_lottype,f_lotbigclass
           from t_base_lotclass
           where f_classsort=1
           )c on a.f_lotid=c.f_lottype
          group by a.f_yearweek,c.f_lotbigclass
         
         
         
         select a.f_yearweek,a.f_lotbigclass,ExpeNum ,locatesalay,countryBonus
          from
          (
              select TO_CHAR(f_DATE,'YYYYMMWW') as f_yearweek ,f_lotbigclass,
                      count(distinct F_Expect) ExpeNum, sum(F_BuyMoney) locatesalay
              from T_Base_User_Buy_Info  
              inner join T_base_Flat_Type t on f_buymodename=F_BuyChannel and F_DEPARTMENT=1 and f_flatname='主站'
              inner join ( select distinct f_lotbigclass,f_lottype from t_base_lotclass t where f_classsort=1) C
              on F_LotID=c.f_lottype
              where f_date between to_date('2007-01-01 00:00:00','yyyy-mm-dd hh24:mi:ss') and to_date('2007-01-10 00:00:00','yyyy-mm-dd hh24:mi:ss')
              group by TO_CHAR(f_DATE,'YYYYMMWW'),f_lotbigclass
          ) a
          left join
          (
              select TO_CHAR(f_date,'YYYYMMWW') as f_yearweek ,f_lotbigclass,sum(F_TaxBefforBonus) countryBonus
              from T_Base_Lot_Send_Bonus s
              inner join T_base_Flat_Type t on s.f_buymodeid=t.f_buymodeid and F_DEPARTMENT=1 and f_flatname='主站'
              inner join ( select distinct f_lotbigclass,f_lottype from t_base_lotclass t where f_classsort=1) C
              on F_LotID=c.f_lottype             
              where f_date between to_date('2007-01-01 00:00:00','yyyy-mm-dd hh24:mi:ss') and to_date('2007-01-10 00:00:00','yyyy-mm-dd hh24:mi:ss')         
              group by TO_CHAR(f_date,'YYYYMMWW'),f_lotbigclass
          )b  on a.f_yearweek=b.f_yearweek and a.f_lotbigclass=b.f_lotbigclass
         
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值