SQL Server存储过程可插入所选数据

Question

I am new to stored procedures to bear with me. In a single procedure I am trying to select data from a table and insert that data into another table. For example here is my selection statement:

SELECT SUM((qty_invoiced * price)) FROM table1,table2
WHERE table1.co_num = table2.co_num 
AND UPPER(table2.strIDX) = 'S' 
AND DATEPART(wk,table1.order_date) = DATEPART(wk,Getdate()) 
AND YEAR(table1.order_date) = YEAR(Getdate()) 

I then want to insert the value of the sum field into another table.

How would I go about doing that in the most efficient way with SQL Server. I can do it in .net code (C# , VB etc) but wanted to do it more efficiently with a stored procedure

Answer 1

You can do this in several ways. Are you expecting the query to return a single row, or multiple rows though?

If you are expecting it to return a single row, perhaps the most straight-forward way would be to save the value into some variable, like this...

DECLARE @mySum DECIMAL

SELECT @mySum = SUM((qty_invoiced * price)) 
FROM table1,table2
WHERE table1.co_num = table2.co_num 
AND UPPER(table2.strIDX) = 'S' 
AND DATEPART(wk,table1.order_date) = DATEPART(wk,Getdate()) 
AND YEAR(table1.order_date) = YEAR(Getdate()) 

And then you can use @mySum in your insert statement as the value to insert. That of course only works if you have a single value returned from the query though.

If you want to insert multiple rows at once from the results of that query, you'd do something like this...

INSERT INTO SomeTable
( TheSumColumn )
SELECT SUM((qty_invoiced * price)) 
FROM table1,table2
WHERE table1.co_num = table2.co_num 
AND UPPER(table2.strIDX) = 'S' 
AND DATEPART(wk,table1.order_date) = DATEPART(wk,Getdate()) 
AND YEAR(table1.order_date) = YEAR(Getdate()) 

Or you can use a SELECT INTO clause. Look at " Adding Rows by Using INSERT and SELECT " for examples of all of these.

Answer 2

This sounds like what you really need is a view.

CREATE VIEW vWeeklySum
as
SELECT 
YEAR(table1.order_date) as [Year],
DATEPART(wk,table1.order_date) as [WeekNo],
SUM(qty_invoiced * price) [Total]
FROM table1
JOIN table2
ON  table1.co_num = table2.co_num 
WHERE UPPER(table2.strIDX) = 'S' 
GROUP BY YEAR(table1.order_date),
DATEPART(wk,table1.order_date)

This way if your raw data changes your view will change without you needing to do anything. Then you could use it like:

SELECT [Total]
FROM vWeeklySum
WHERE [Year] = 2015
AND [WeekNo] = 5

If you still want to save it to a table (a snapshot), your stored proc would look something like:

CREATE PROC SaveWeek(DateTime @Date)
AS

DELETE FROM myTable
WHERE WeekNo = DATEPART(wk,@Date) 
AND [Year] = YEAR(@Date)

INSERT INTO myTable
([Year], WeekNo, [Total])
SELECT 
YEAR(@Date),
DATEPART(wk,@Date),
SUM(qty_invoiced * price)
FROM table1
JOIN table2
ON  table1.co_num = table2.co_num 
WHERE UPPER(table2.strIDX) = 'S' 
AND DATEPART(wk,table1.order_date) = DATEPART(wk,@Date) 
AND YEAR(table1.order_date) = YEAR(@Date)
GROUP BY YEAR(@Date),
DATEPART(wk,@Date)

Then you could call it with

DateTime @now
set @now = getdate()
exec SaveWeek @now

Answer 3

If you want to be efficient, you should never use functions for the SARG fields, so you should do something like this:

INSERT INTO table3
SELECT SUM((qty_invoiced * price))
FROM table1 t1,table2 t2
WHERE t1.co_num = t2.co_num 
AND t2.strIDX in ('s','S') 
AND t1.order_date >= @startdate 
AND t1.order_date < @enddate

Calculate the correct values to the variables before the statement. @enddate is intentionally the next day, just in case your field is datetime.