请帮忙看看该如何优化sql

蓝海豚37

数据源表是详单表bill_call_ticket，插入数据中间表report_cell_day。详单表按天分区，问题是我用这个sql统计一般只要10分种可以统计出来，可是昨天统计却花了9个小时，很是纳闷。其中where call_type=0 完全可以去除，因为详单表call_type字段值只有0，只是怕出现非0的状况才加上去的。详单表建有busi_type,ticket_type的索引，但是我看了一下，索引未能用上。请高手指点如何优化！不甚感激！
insert all

when busi_type in ('10','50','54','55','AA','A1','A2','A3','51','52','53','58','56')
     or (user_type='1' and ticket_type='13') or (busi_type='30' and ticket_type in ('AA','AB')) then
into report_cell_day
values(account_month,day_id,USER_TYPE,0,
       ticket_type,BUSI_TYPE,roam_type,
       FEE_TYPE,holiday_type,PHS_own_ZONE_CODE ,PHS_CUR_ZONE_CODE ,
       AREA_ID,group_id,other_zone_code,OUT_PARTNER_ID ,
       interconnect_type,interconnect_code,
       in_trunk_id,out_trunk_id,machine_name,disct_type,
       duration,fee_disct,
       charge,charge_counts,fee_other,fee_other_counts,fee_sum,
       sum_tickets ,charge_duration,fee_other_duration)

when busi_type in ('20','40','41','42','43','44','45','46','47','48','49','70','12','13')
     or ticket_type='12' or (user_type='0' and ticket_type='13') or (busi_type='30' and ticket_type='AC') then
into report_cell_day
values(account_month,day_id,USER_TYPE,0,
       '10'||busi_type||ticket_type,'10',roam_type,
       FEE_TYPE,holiday_type,PHS_own_ZONE_CODE ,PHS_CUR_ZONE_CODE ,
       AREA_ID,group_id,other_zone_code,OUT_PARTNER_ID ,
       interconnect_type,interconnect_code,
       in_trunk_id,out_trunk_id,machine_name,disct_type,
       duration,fee_disct,
       charge,charge_counts,0,0,charge,
       sum_tickets ,charge_duration,0)

when busi_type in ('20','40','41','42','43','44','45','46','47','48','49','70','12','13')
     or ticket_type='12' or (user_type='0' and ticket_type='13') or (busi_type='30' and ticket_type='AC') then
into report_cell_day
values(account_month,day_id,USER_TYPE,0,
       ticket_type,BUSI_TYPE,roam_type,
       FEE_TYPE,holiday_type,PHS_own_ZONE_CODE ,PHS_CUR_ZONE_CODE ,
       AREA_ID,group_id,other_zone_code,OUT_PARTNER_ID ,
       interconnect_type,interconnect_code,
       in_trunk_id,out_trunk_id,machine_name,disct_type,
       duration,fee_disct,
       0,0,fee_other,fee_other_counts,fee_other,
       sum_tickets ,0,fee_other_duration)      

select account_month,partitionx day_id,USER_TYPE,0,
       ticket_type,BUSI_TYPE,roam_type,
       null FEE_TYPE,substr(FEE_TYPE,-1,1) holiday_type,nvl(PHS_own_ZONE_CODE,'111') PHS_own_ZONE_CODE,
       nvl(PHS_CUR_ZONE_CODE,'111') PHS_CUR_ZONE_CODE,
       nvl(AREA_ID,0) AREA_ID,null group_id,nvl(other_zone_code,'111') other_zone_code,
       nvl(OUT_PARTNER_ID,'X') OUT_PARTNER_ID,
       null interconnect_type,NULL interconnect_code,
       null in_trunk_id,decode(user_type,'0',out_trunk_id,'0') out_trunk_id,null machine_name,null disct_type,
       sum(decode(BUSI_TYPE,'10',duration,'20',duration,'30',duration,'60',duration,'12',duration,'13',duration,0)) duration,
       sum(fee_disct) fee_disct,
       sum(nvl(charge,0)) charge,
       sum(case when nvl(charge,0)=0 then 0 when busi_type not in ('10','20','30','60','12','13') then 0 else charge_counts end ) charge_counts,
       sum(nvl(fee_other,0)) fee_other,
       sum(case when nvl(fee_other,0)=0 then 0 when busi_type not in ('10','20','30','60','12','13') then 0 else fee_other_counts end) fee_other_counts,
       sum(fee_sum) fee_sum,
       count(*) sum_tickets ,
       sum(case when nvl(charge,0)=0 then 0 when busi_type not in ('10','20','30','60','12','13') then 0 else bsc_bill_duration end) charge_duration,
       sum(case when nvl(fee_other,0)=0 then 0 when busi_type not in ('10','20','30','60','12','13') then 0 else oth_bill_duration end) fee_other_duration

from bill_call_ticket partition(b20070301)  

where call_type='0'   
group by account_month,partitionx, USER_TYPE,ticket_type,BUSI_TYPE,
        roam_type,substr(FEE_TYPE,-1,1),phs_own_zone_code,PHS_CUR_ZONE_CODE,AREA_ID,other_zone_code,
       OUT_PARTNER_ID,out_trunk_id

gongzidong

哇 看着头晕！不想动脑子想了。
只能提醒句，建中间表吧。