导出 SQL 查询疑难解答

Question

我已经编写了一个查询，并且能够成功运行它。当我尝试导出它时，它的导出速度非常慢，即每 10-15 分钟 50 行，我设法在 10 小时内导出了大约 50K 数据。如果此查询有任何问题，请帮助我。我正在使用 sql 开发人员

谢谢

我想要的结果是按列显示的：

|GLUSR_USR_CUSTTYPE_NAME|GLUSR_USR_ID|S_COUNT|MAX_DATE as Last Meeting|Min_DATE as first Meeting|Total Meeting in Last 90 Days|Total Meeting in last 180 days| active in last 30 Days| active in last 60 Days| active in last 90 Days|

sql:

Select GLUSR_USR_ID,
    GLUSR_USR_CUSTTYPE_NAME,
    ( select  Count(1)                   from STS_DSR_SALES@mainr,STS_Company@mainr where STS_DSR_SALES.FK_STS_COMPANY_ID=CompanyID and GLUSR_USR_ID=STS_FK_GLUSR_ID)S_COUNT,
    (Select MAX(STS_DSR_SALES_DATE)      from STS_DSR_SALES@mainr,STS_Company@mainr where STS_DSR_SALES.FK_STS_COMPANY_ID=COMPANYID)MAX_DATE,
    (Select min(STS_DSR_SALES_DATE)      from STS_DSR_SALES@mainr,STS_Company@mainr where STS_DSR_SALES.FK_STS_COMPANY_ID=COMPANYID)Min_DATE,
    (Select count (1) STS_DSR_SALES_DATE from STS_DSR_SALES@mainr,STS_COMPANY@mainr where STS_DSR_SALES.FK_STS_COMPANY_ID= COMPANYID and GLUSR_USR_ID=STS_FK_GLUSR_ID and trunc(STS_DSR_SALES_DATE)>SYSDATE-90)last_90,
    (Select count (1) STS_DSR_SALES_DATE from STS_DSR_SALES@mainr,STS_COMPANY@mainr where STS_DSR_SALES.FK_STS_COMPANY_ID= COMPANYID and GLUSR_USR_ID=STS_FK_GLUSR_ID and trunc(STS_DSR_SALES_DATE)>SYSDATE-180)last_180,
    (Select Count (1) STS_DSR_SALES_DATE from STS_DSR_SALES@mainr,GLUSR_CLCKSTRM_4HOTLEAD_ARCH@mainr where GLUSR_USR_ID=FK_GLUSR_USR_ID and STS_DSR_SALES_DATE=Report_DATE and TRUNC(REPORT_DATE)>SYSDATE-30)ACTIVE_30,
    (Select Count (1) STS_DSR_SALES_DATE from STS_DSR_SALES@mainr,GLUSR_CLCKSTRM_4HOTLEAD_ARCH@mainr where GLUSR_USR_ID=FK_GLUSR_USR_ID and STS_DSR_SALES_DATE=Report_DATE and TRUNC(REPORT_DATE)>SYSDATE-60)ACTIVE_60,
    (Select Count (1) STS_DSR_SALES_DATE from STS_DSR_SALES@mainr,GLUSR_CLCKSTRM_4HOTLEAD_ARCH@mainr where GLUSR_USR_ID=FK_GLUSR_USR_ID and STS_DSR_SALES_DATE=Report_DATE and TRUNC(REPORT_DATE)>SYSDATE-90)ACTIVE_90
 From
    (
        select GLUSR_USR_ID, fk_sts_company_id,GLUSR_USR_CUSTTYPE_NAME 
        from STS_DSR_SALES@mainr,GLUSR_USR@mainr,STS_COMPANY@MAINR         
        where STS_FK_GLUSR_ID=GLUSR_USR_ID
        and fk_sts_company_id=CompanyID
        and    trunc(sts_dsr_sales_date) between '01-oCT-19' and '31-oCT-19'
    )  

Answer 1

为了获得答案/结果，您访问表格的次数过多。 尝试使用 WITH() 子句构建内联查询，并一次一个地逐步构建数据管道。

With Q1 as ( 
inline_query on 2 tables STS_DSR_SALES and STS_Company which defines the overall scope of your query/ask 
),
Q2 as (
select ... <All/90/180 day aggregates defined via sum(case statements)>
from Q1 inner join <other_table T3 GLUSR_USR>
where ... (keys)
),
Q3 as (
select ... <All/30/60/90 day aggregates defined via sum(case statements)>
from Q1 inner join <other_table T4 GLUSR_CLCKSTRM_4HOTLEAD_ARCH>
where ... (keys)
)
--final result
select * from (
select * from Q2 inner join Q3 on ...
)
pivot ( 
-- convert rows to cols ... get your KPIs to line up as columns 
);

Answer 2

很难了解全貌，因为您指的是没有表名/别名的列，并且依赖于这些确切的名称仅出现在其中一个表中，而且您正在从 select 子句中进行选择。

您的查询似乎有很多具有最小/最大或不同日期范围的重复部分。我很确定这些可以使用case 语句组合成较少数量的查询，比如

sum(case when TRUNC(REPORT_DATE)>SYSDATE-30 then 1 end),
sum(case when TRUNC(REPORT_DATE)>SYSDATE-60 then 1 end),
sum(case when TRUNC(REPORT_DATE)>SYSDATE-90 then 1 end),

所以你会摆脱一些查询。

不再记得这是否是一件坏事（或仅在 SQL Server 中），而且自从我上次使用 Oracle 以来已经有十多年了，所以事情可能也发生了变化 :)

但您可能想测试替换它：

trunc(sts_dsr_sales_date) between '01-oCT-19' and '31-oCT-19'

用这样的东西，你就可以摆脱这个功能：

sts_dsr_sales_date >= '01-OCT-19' and sts_dsr_sales_date < '1-NOV-19'