有关SQL Server中的联接的查询

Question

I tried the following code

SELECT
 a.sodtl_order_no  'Document_Number',
 a.ATTDESC   'Attribute_Desc'
 ,b.Customer_name 'Cust_Name'
 ,b.Address_Line1   'Add_Line1'
 ,b.City  'City'
 ,b.State 'State'
 ,b.Zip_Code 'Zipcode'
 ,b.Item_Description 'Itemdescription'
 ,b.Document_Date
,convert(int,c.sohdr_total_value)'Total_Amount'
from  QBE0017_MaintainSaleOrder_QualityAttributes_VW a INNER JOIN
 QBE0022_SALEORDER_VW b  ON a.sodtl_order_no = b.Document_number   AND
 a.sodtl_line_no=b.line_no
 INNER JOIN QBE0017_MaintainSaleOrder_Hdr_VW c ON
 b.Document_Number=c.sohdr_order_no
where b.LINE_NO IN(select top 10 d.sodtl_line_no from QBE0017_MaintainSaleOrder_Dtl_VW   d INNER JOIN  QBE0017_SO_TCD_VW e ON d.sodtl_line_no=e.[LINENO]
)

I have used inner query and I need to display some columns from d table and some from e table ie where should I mention the column name that needs to be displayed ?

Answer 1

SELECT
 a.sodtl_order_no  'Document_Number',
 a.ATTDESC   'Attribute_Desc'
 ,b.Customer_name 'Cust_Name'
 ,b.Address_Line1   'Add_Line1'
 ,b.City  'City'
 ,b.State 'State'
 ,b.Zip_Code 'Zipcode'
 ,b.Item_Description 'Itemdescription'
 ,b.Document_Date
,convert(int,c.sohdr_total_value)'Total_Amount'
,f.columnOne
,f.columnTwo
from  QBE0017_MaintainSaleOrder_QualityAttributes_VW a INNER JOIN
 QBE0022_SALEORDER_VW b  ON a.sodtl_order_no = b.Document_number   AND
 a.sodtl_line_no=b.line_no
 INNER JOIN QBE0017_MaintainSaleOrder_Hdr_VW c ON
 b.Document_Number=c.sohdr_order_no
INNER JOIN (select top 10 d.sodtl_line_no, d.columnOne, e.columnTwo from QBE0017_MaintainSaleOrder_Dtl_VW   d INNER JOIN  QBE0017_SO_TCD_VW e ON d.sodtl_line_no=e.[LINENO]
) f ON b.LINE_NO = f.sodtl_line_no

As long as you select out the columns from you sub-select you can use them in your outer select that you join on to.

Note: If you're using a TOP clause you should probably also be using an ORDER BY clause to guarantee the order in which you get the results as SQL can decide on any old order if you don't explicitly tell it what you want which can lead to inconsistent results.

Answer 2

SELECT
a.sodtl_order_no  AS Document_Number,
a.ATTDESC   AS Attribute_Desc
,b.Customer_name AS Cust_Name
,b.Address_Line1   AS Add_Line1
,b.City  AS City
,b.State AS State
,b.Zip_Code AS Zipcode
,b.Item_Description AS Itemdescription
,b.Document_Date
,convert(int,c.sohdr_total_value) AS Total_Amount
from  QBE0017_MaintainSaleOrder_QualityAttributes_VW a INNER JOIN
QBE0022_SALEORDER_VW b  ON a.sodtl_order_no = b.Document_number   AND
a.sodtl_line_no=b.line_no
INNER JOIN QBE0017_MaintainSaleOrder_Hdr_VW c ON
b.Document_Number=c.sohdr_order_no
where b.LINE_NO IN(select top 10 d.sodtl_line_no from QBE0017_MaintainSaleOrder_Dtl_VW   d INNER JOIN  QBE0017_SO_TCD_VW e ON d.sodtl_line_no=e.[LINENO])

Answer 3

You could replace this subselection in the WHERE :

where b.LINE_NO IN(select top 10 d.sodtl_line_no from QBE0017_MaintainSaleOrder_Dtl_VW   d INNER JOIN  QBE0017_SO_TCD_VW e ON d.sodtl_line_no=e.[LINENO]
)

with two extra inner joins, like this:

SELECT  a.sodtl_order_no AS Document_Number
        ,a.ATTDESC AS Attribute_Desc
        ,b.Customer_name AS Cust_Name
        ,b.Address_Line1 AS Add_Line1
        ,b.City AS City
        ,b.State AS [State]
        ,b.Zip_Code AS Zipcode
        ,b.Item_Description AS Itemdescription
        ,b.Document_Date
        ,CONVERT(INT,c.sohdr_total_value) AS Total_Amount
        ,d.YYY -- values from d
        ,e.ZZZ -- values from e
FROM    QBE0017_MaintainSaleOrder_QualityAttributes_VW a 
INNER JOIN QBE0022_SALEORDER_VW b ON a.sodtl_order_no = b.Document_number AND a.sodtl_line_no = b.line_no
INNER JOIN QBE0017_MaintainSaleOrder_Hdr_VW c ON b.Document_Number = c.sohdr_order_no
INNER JOIN QBE0017_MaintainSaleOrder_Dtl_VW d ON d.sodtl_line_no = b.LINE_NO
INNER JOIN QBE0017_SO_TCD_VW e ON d.sodtl_line_no = e.[LINENO];

The only thing you're losing then is the subselect TOP 10 filter. I think you could somehow put this in your query again, but I doubt if you really want that. It filters your data in a weird way, saying "Give me 10 random values back, I don't mind what they are, and I'll use that data from now on". You can probably think of some better filters to put in :-)