我会在等式中引入一个表格UDF，如清单C所示。注意，我可以用两种方法调用函数：提交 CustomerID将表格限定为那个顾客的订单，或提交一个NULL获得所有顾客订单清单。如果我提交CustomerID，那么订单清单就出现在 OrderDate中；如果我提交NULL，清单就被分组并在CustomerID中由OrderDate排序。
set ANSI_NULLS ON
set QUOTED_IDENTIFIER ON
go
-- =============================================
-- Author: Arthur Fuller
-- Create date: 23 Aug 2006
-- Description: Table UDF to return Order Totals by Customer
-- Example: SELECT * FROM dbo.CustomerOrderTotals_fnt("VINET")
-- Example: SELECT * FROM dbo.CustomerOrderTotals_fnt(NULL)
-- Notes: This udf is designed to serve two
-- purposes. Pass a CustomerID to limit the rows to
-- that customer, or pass nothing to get all customers
-- =============================================
ALTERFUNCTION [dbo].[CustomerOrderTotals_fnt]
(
-- Add the parameters for the function here
@CustomerID varchar(5)=NULL
)
RETURNS TABLE
AS
RETURN
(
-- Add the SELECT statement with parameter references here
SELECTTOP 100 PERCENT
dbo.Customers.CustomerID,
dbo.Customers.CompanyName,
dbo.Orders.OrderID,
dbo.Orders.OrderDate,
dbo.OrderDetailsSumByOrderID_vue.TotalAmount
FROM
dbo.Customers
INNERJOIN
dbo.Orders ON dbo.Customers.CustomerID = dbo.Orders.CustomerID
INNERJOIN
dbo.OrderDetailsSumByOrderID_vue
ON dbo.Orders.OrderID = dbo.OrderDetailsSumByOrderID_vue.OrderID
WHERE
dbo.Customers.CustomerID = @CustomerID
OR @CustomerID ISNULL
ORDERBY dbo.Orders.OrderDate
)
--try it with these:
--SELECT * FROM dbo.CustomerOrderTotals_fnt("VINET")
--SELECT * FROM dbo.CustomerOrderTotals_fnt(NULL) |
清单C
-- ================================================
-- Template generated from Template Explorer using:
-- Create Inline Function (New Menu).SQL
--
-- Use the Specify Values for Template Parameters
-- command (Ctrl-Shift-M) to fill in the parameter
-- values below.
--
-- This block of comments will not be included in
-- the definition of the function.
-- ================================================
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
-- =============================================
-- Author: Arthur Fuller
-- Create date: 23 Aug 2006
-- Description: Return total sales by Customer
-- Example:
-- SELECT CustomerID, CustomerTotal
-- FROM DBO.CustomerGrandTotal_fnt(null)
-- SELECT CustomerID, CustomerTotal
-- FROM DBO.CustomerGrandTotal_fnt("VINET")
-- SELECT CustomerID, CustomerTotal
-- FROM DBO.CustomerGrandTotal_fnt("VINET")
-- =============================================
CREATEFUNCTION CustomerGrandTotal_fnt
(
@CustomerID varchar(5)
)
RETURNS TABLE
AS
RETURN
(
-- Add the SELECT statement with parameter references here
SELECTTOP 100 PERCENT
CustomerID,
SUM(TotalAmount)AS CustomerTotal
FROM dbo.CustomerOrderTotals_fnt(@CustomerID)
WHERE dbo.CustomerOrderTotals_fnt.CustomerID = @CustomerID
OR @CustomerID ISNULL
GROUPBY CustomerID
ORDERBY CustomerID
)
GO |
清单D
更深入一步，我每个顾客打印一行，用一列显示每名顾客的总销售额。查看清单D。并不是所有项目都得到组合。在最外层，我能够以两种方式调用这个函数：提交合法CustomerID或UNLL。以下是它们各自的实例：
SELECT CustomerID, CustomerTotal
FROM DBO.CustomerGrandTotal_fnt(null)
SELECT CustomerID, CustomerTotal
FROM DBO.CustomerGrandTotal_fnt("VINET") |
现在你有了一组精确的工具来帮助你挖掘适当层次的细节。计算列OrderDetails.ExtendedAmount隐藏了一个复杂层次，从那我又深入到我希望隐藏的细节。这就是我把这些对象叫做原子和分子查询的原因。我使用“查询”一词来概括这样一个事实，即讨论的对象是视图还是表格UDF并不重要。（虽然由于命名规则，suffix _fnt表示一个表格UDF，而suffix _vue表示一个视图。）
必须承认，Northwind数据库中没有大量的行，但我认为这种技巧可以得到很好地扩充。更为重要的是，我喜欢它提供的粒度。我能够做我想做的，并按要求重新使用或重新组合原子。
