Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

How can I select multiple columns from a subquery (in SQL Server) that should have one record (select top 1) for each record in the main query? [duplicate]

Tags:

sql

People also ask

When a subquery used multiple columns of data with one or more rows?

Multiple row subquery returns one or more rows to the outer SQL statement. You may use the IN, ANY, or ALL operator in outer query to handle a subquery that returns multiple rows. Contents: Using IN operator with a Multiple Row Subquery.


Here's generally how to select multiple columns from a subquery:

SELECT
     A.SalesOrderID,
     A.OrderDate,
     SQ.Max_Foo,
     SQ.Max_Foo2
FROM
     A
LEFT OUTER JOIN
     (
     SELECT
          B.SalesOrderID,
          MAX(B.Foo) AS Max_Foo,
          MAX(B.Foo2) AS Max_Foo2
     FROM
          B
     GROUP BY
          B.SalesOrderID
     ) AS SQ ON SQ.SalesOrderID = A.SalesOrderID

If what you're ultimately trying to do is get the values from the row with the highest value for Foo (rather than the max of Foo and the max of Foo2 - which is NOT the same thing) then the following will usually work better than a subquery:

SELECT
     A.SalesOrderID,
     A.OrderDate,
     B1.Foo,
     B1.Foo2
FROM
     A
LEFT OUTER JOIN B AS B1 ON
     B1.SalesOrderID = A.SalesOrderID
LEFT OUTER JOIN B AS B2 ON
     B2.SalesOrderID = A.SalesOrderID AND
     B2.Foo > B1.Foo
WHERE
     B2.SalesOrderID IS NULL

You're basically saying, give me the row from B where I can't find any other row from B with the same SalesOrderID and a greater Foo.


SELECT a.salesorderid, a.orderdate, s.orderdate, s.salesorderid
FROM A a
OUTER APPLY (SELECT top(1) *
             FROM B b WHERE a.salesorderid = b.salesorderid) as s
WHERE A.Date BETWEEN '2000-1-4' AND '2010-1-4'

You'll have to make a join:

SELECT A.SalesOrderID, B.Foo
FROM A
JOIN B bo ON bo.id = (
     SELECT TOP 1 id
     FROM B bi
     WHERE bi.SalesOrderID = a.SalesOrderID
     ORDER BY bi.whatever
     )
WHERE A.Date BETWEEN '2000-1-4' AND '2010-1-4'

, assuming that b.id is a PRIMARY KEY on B

In MS SQL 2005 and higher you may use this syntax:

SELECT SalesOrderID, Foo
FROM (
  SELECT A.SalesOrderId, B.Foo,
         ROW_NUMBER() OVER (PARTITION BY B.SalesOrderId ORDER BY B.whatever) AS rn
  FROM A
  JOIN B ON B.SalesOrderID = A.SalesOrderID
  WHERE A.Date BETWEEN '2000-1-4' AND '2010-1-4'
) i
WHERE rn

This will select exactly one record from B for each SalesOrderId.


select t1.*, sq.*
from table1 t1,
   (select a,b,c from table2 ...) sq
where ...

i think that is what you want.

SELECT 
      A.SalesOrderID, 
      A.OrderDate, 
      FooFromB.*

FROM A,
     (SELECT TOP 1 B.Foo
      FROM B
      WHERE A.SalesOrderID = B.SalesOrderID
      ) AS FooFromB

WHERE A.Date BETWEEN '2000-1-4' AND '2010-1-4'