MDAC 2.5 SDK - OLE DB Programmer's Reference
Chapter 26: Mapping MDX to SQL Statements
The COLUMNS axis expression consists of the CROSSJOIN function applied on two literal sets. Applying the mappings described in "Literal Sets" generates two virtual tables. These virtual tables are used as input to the SQL statements described in the section "CROSSJOIN Function."
The result of these operations yields the following two tables (denoted as Table C and Table CPrime):
Name1 | Name2 | Rank |
Products.[All].Office.Access | Geography.[All].USA.Kansas | 1 |
Products.[All].Office.Access | Geography.[All].USA.NewYork.Buffalo | 2 |
Products.[All].Office.Access | Geography.[All].USA.Kansas.Topeka | 3 |
Products.[All].Office.Access | Geography.[All].USA | 4 |
Products.[All].Office.Access | Geography.[All].Canada | 5 |
Products.[All].Office.Word | Geography.[All].USA.Kansas | 6 |
Products.[All].Office.Word | Geography.[All].USA.NewYork.Buffalo | 7 |
Products.[All].Office.Word | Geography.[All].USA.Kansas.Topeka | 8 |
Products.[All].Office.Word | Geography.[All].USA | 9 |
Products.[All].Office.Word | Geography.[All].Canada | 10 |
Product | ProdLine | Product.[All] | City | State | Country | Geography.[All] | Rank |
Products.[All].Office.Access | Products.[All].Office | Product.[All] | ALL() | Geography.USA.Kansas | Geography.USA | Geography.[All] | 1 |
Products.[All].Office.Access | Products.[All].Office | Product.[All] | Geography.USA.NewYork.Buffalo | Geography.USA.NewYork | Geography.USA | Geography.[All] | 2 |
Products.[All].Office.Access | Products.[All].Office | Product.[All] | Geography.USA.Kansas.Topeka | Geography.USA.Kansas | Geography.USA | Geography.[All] | 3 |
Products.[All].Office.Access | Products.[All].Office | Product.[All] | ALL() | ALL() | Geography.USA | Geography.[All] | 4 |
Products.[All].Office.Access | Products.[All].Office | Product.[All] | ALL() | ALL() | Geography.Canada | Geography.[All] | 5 |
Products.[All].Office.Word | Products.[All].Office | Product.[All] | ALL() | Geography.USA.Kansas | Geography.USA | Geography.[All] | 6 |
Products.[All].Office.Word | Products.[All].Office | Product.[All] | Geography.USA.NewYork.Buffalo | Geography.USA.NewYork | Geography.USA | Geography.[All] | 7 |
Products.[All].Office.Word | Products.[All].Office | Product.[All] | Geography.USA.Kansas.Topeka | Geography.USA.Kansas | Geography.USA | Geography.[All] | 8 |
Products.[All].Office.Word | Products.[All].Office | Product.[All] | ALL() | ALL() | Geography.USA | Geography.[All] | 9 |
Products.[All].Office.Word | Products.[All].Office | Product.[All] | ALL() | ALL() | Geography.Canada | Geography.[All] | 10 |
Cprime is obtained from C by using the following steps:
SELECT COUNT DISTINCT COMPONENT(Name, -1) FROM C
In the current example, this is 2.
SELECT DISTINCT COMPONENT(Name, -1) FROM C
In the current example, this is Products, Geography.
SELECT MAX(LEVEL(Name)) FROM C GROUP BY COMPONENT(Name, -1)
In the current example, this is 3 for Geography and 2 for Products.
CREATE LOCAL TEMPORARY VIEW C1(Name, Rank) AS
SELECT Name, NewRank AS Rank
FROM
(SELECT Name1 FROM C WHERE COMPONENT(Name, -1) = "Products"
RANK ROWS AS NewRank RANKORDER BY Rank)
CREATE LOCAL TEMPORARY VIEW C2(Name, Rank) AS
SELECT Name, NewRank AS Rank
FROM
(SELECT Name2 FROM C WHERE COMPONENT(Name, -1) = "Geography"
RANK ROWS AS NewRank RANKORDER BY Rank)
SELECT Product AS COMPONENT(C1.Name, 2),
ProdLine AS COMPONENT(C1.Name, 1),
Product.[All] AS COMPONENT(C1.Name, 0),
City AS COMPONENT(C2.Name, 3),
State AS COMPONENT(C2.Name, 2),
Country AS COMPONENT(C2.Name, 1),
Geography.[All] AS COMPONENT(C2.Name, 0)
NewRank AS Rank
FROM
(C1 CROSS JOIN C2) RANK ROWS AS NewRank
RANKORDER BY C1.Rank, C2.Rank