Question

我有两个表，t1，t2在架构中 - ＆＃39; Schema， t1有100万行，大约有30列。

我正在运行如下所示的查询，

select b.Contract_ID,
b.Signed, 
b.Funding_Source,
b.Theater, b.Country,
b.Partner, b.BEID, 
b.Contract_Type, b.TCV,
b.Technology, 
b.Partner_Signature_Date,
b.Signature_Date, b.Term,
b.Purchase_Order, 
b.Contract_Baseline,
b.Organic_Rate, b.Target_Organic, 
b.Target_Incremental,
b.Total_Bookings_Target,
b.Tracking_Start, 
b.Tracking_End, b.Qualifier_BEID, 
b.Qualifier_BEGeoID, 
a.`CSC Global Ultimate ID`, 
a.`CSC Company Target ID`,
b.Qualifier_Theater, 
b.Qualifier_Country, a.SCMS,
b.Qualifier_POType, b.MCO, 
b.Contract_Status, b.Accrual_Start, 
b.Accrual_End, b.Booking_Start, 
b.Booking_End, b.TMS_Level_1, 
b.TMS_Level_2, b.TMS_Level_3, 
b.Pf, 
b.Allocation,a.`Fiscal Period ID`,
a.Bookings*b.Allocation as Bookings  
from
(SELECT * FROM `Schema`.`t1` where `BEID` in 
(select distinct (BEID) from Schema.`t2`)) a
INNER JOIN
(SELECT * FROM `Schema`.`t2`) b
where
a.`Business Entity Id` = b.Qualifier_BEID
and a.`BE Geo Id` = b.Qualifier_BEGeoID
and a.`Country` = b.Qualifier_Country
and a.`PO Type` = b.Qualifier_POType 
and a.`Product Family` = b.Pf

这是查询，但需要45分钟才能运行。关于一些事情的一点背景。

select distinct (BEID) from Schema. t2

产生137个数值（最多4个数字）。

我已将BIED列编入索引，因此别名＆＃39; a＆＃39;计算起来要快一点。

Answer 1

首先，写下这样的查询

select b.Contract_ID,

b.Signed, 
b.Funding_Source,
b.Theater, b.Country,
b.Partner, b.BEID, 
b.Contract_Type, b.TCV,
b.Technology, 
b.Partner_Signature_Date,
b.Signature_Date, b.Term,
b.Purchase_Order, 
b.Contract_Baseline,
b.Organic_Rate, b.Target_Organic, 
b.Target_Incremental,
b.Total_Bookings_Target,
b.Tracking_Start, 
b.Tracking_End, b.Qualifier_BEID, 
b.Qualifier_BEGeoID, 
a.`CSC Global Ultimate ID`, 
a.`CSC Company Target ID`,
b.Qualifier_Theater, 
b.Qualifier_Country, a.SCMS,
b.Qualifier_POType, b.MCO, 
b.Contract_Status, b.Accrual_Start, 
b.Accrual_End, b.Booking_Start, 
b.Booking_End, b.TMS_Level_1, 
b.TMS_Level_2, b.TMS_Level_3, 
b.Pf, 
b.Allocation,a.`Fiscal Period ID`,
a.Bookings*b.Allocation as Bookings



from

`Schema`.`t1` a
INNER JOIN
`Schema`.`t2` b on a.BEID = b.BEID


where
a.`Business Entity Id` = b.Qualifier_BEID

and a.`BE Geo Id` = b.Qualifier_BEGeoID 

and a.`Country` = b.Qualifier_Country 

and a.`PO Type` = b.Qualifier_POType 

and a.`Product Family` = b.Pf

第二，确保您在Schema。t1上创建了一个索引。并且与schema.t2一样

Answer 2

使其成为没有子查询的标准连接，并将所有条件移至连接条件。仅当BEID在t2中多次出现时才使用distinct。

select distinct
    -- various columns as per your posted query
from t1 a
join t2 b on a.BEID = b.BEID
    and a.`Business Entity Id` = b.Qualifier_BEID
    and a.`BE Geo Id` = b.Qualifier_BEGeoID
    and a.`Country` = b.Qualifier_Country
    and a.`PO Type` = b.Qualifier_POType 
    and a.`Product Family` = b.Pf

将索引放在t2(BEID)上，或者更好地放在t2(BEID, Qualifier_BEID, Qualifier_BEGeoID)

上

我省略了这些列，因为它们对问题并不重要。

百万行的mysql where子句

2 个答案: