Из описания MDX следует (будь я трижды не прав), что
NonEmtpyCrossJoin(
{Dim1.Member1:Dim1.MemberN},
{Dim2.Member1:Dim2.MemberM},
{Dim2.Member1:Dim2.MemberK}, 1)
эквивалентно
Extract(
NonEmtpyCrossJoin(
{Dim1.Member1:Dim1.MemberN},
{Dim2.Member1:Dim2.MemberM},
{Dim2.Member1:Dim2.MemberK})
Dim1)
только второй вариант требует больше ресурсов.
но как показывает практика это не так.
В качестве подтверждения два MDX запроса к FoodMart 2000, которые возвращают разный результат.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
with
set [ProductSet] as ' {[Product].[All Products].[Food].[Baking Goods],
descendants([Product].[All Products].[Food].[Baking Goods],
[Product].[Product Name])}'
set [MeasureSet] as '{[Measures].[Sales Count], [Measures].[Store Sales]}'
select
crossjoin([Time].[ 1998 ].children, [MeasureSet]) on columns,
NonEmptyCrossJoin([ProductSet], [MeasureSet], {[Time].[ 1998 ]}, 1 ) on rows
from [Warehouse and Sales]
where [Customers].[All Customers].[USA].[WA].[Anacortes]
и
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
with
set [ProductSet] as ' {[Product].[All Products].[Food].[Baking Goods],
descendants([Product].[All Products].[Food].[Baking Goods],
[Product].[Product Name])}'
set [MeasureSet] as '{[Measures].[Sales Count], [Measures].[Store Sales]}'
select
crossjoin([Time].[ 1998 ].children, [MeasureSet]) on columns,
extract(NonEmptyCrossJoin([ProductSet], [MeasureSet], {[Time].[ 1998 ]}), [Product]) on rows
from [Warehouse and Sales]
where [Customers].[All Customers].[USA].[WA].[Anacortes]
