次の方法で共有


集合関数の使用

集合関数は、ディメンション、階層、レベルからセットを取得します。あるいは、それらのオブジェクト内でのメンバの絶対位置および相対位置をトラバースすることにより、さまざまな方法でセットを構築します。

集合関数は、メンバ関数や組関数と同様に、Analysis Services で使用される多次元構造を操作するために不可欠です。さらに、セット式が多次元式 (MDX) クエリの軸を定義するため、集合関数は MDX クエリから結果を得るうえでも不可欠です。

最も一般的な集合関数の 1 つに Members (セット) (MDX) 関数があります。この関数は、1 つのディメンション、階層、またはレベルのメンバをすべて含むセットを取得します。クエリ内でのこの関数の使用例を次に示します。

SELECT

//Returns all of the members on the Measures dimension

[Measures].MEMBERS

ON Columns,

//Returns all of the members on the Calendar Year level of the Calendar Year Hierarchy

//on the Date dimension

[Date].[Calendar Year].[Calendar Year].MEMBERS

ON Rows

FROM [Adventure Works]

一般的に使用されるもう 1 つの関数として、Crossjoin (MDX) 関数があります。この関数は、パラメータとして渡されるセットのデカルト積を表す組のセットを返します。つまり、この関数を使用すると、クエリに "入れ子になった" 軸または "クロス集計された" 軸を作成できます。

SELECT

//Returns all of the members on the Measures dimension

[Measures].MEMBERS

ON Columns,

//Returns a set containing every combination of all of the members

//on the Calendar Year level of the Calendar Year Hierarchy

//on the Date dimension and all of the members on the Category level

//of the Category hierarchy on the Product dimension

Crossjoin(

[Date].[Calendar Year].[Calendar Year].MEMBERS,

[Product].[Category].[Category].MEMBERS)

ON Rows

FROM [Adventure Works]

Descendants (MDX) 関数は、Children 関数と似ていますがより強力です。この関数は、階層内の 1 つ以上のレベルにある任意のメンバの子孫を返します。

SELECT

[Measures].[Internet Sales Amount]

ON Columns,

//Returns a set containing all of the Dates beneath Calendar Year

//2004 in the Calendar hierarchy of the Date dimension

DESCENDANTS(

[Date].[Calendar].[Calendar Year].&[2004]

, [Date].[Calendar].[Date])

ON Rows

FROM [Adventure Works]

Order (MDX) 関数を使用すると、特定の数値式に従って、セットの内容を昇順または降順で並べ替えることができます。次のクエリでは、上記のクエリと同じ、行のメンバを返しますが、ここでは Internet Sales Amount メジャーでこのメンバを並べ替えます。

SELECT

[Measures].[Internet Sales Amount]

ON Columns,

//Returns a set containing all of the Dates beneath Calendar Year

//2004 in the Calendar hierarchy of the Date dimension

//ordered by Internet Sales Amount

ORDER(

DESCENDANTS(

[Date].[Calendar].[Calendar Year].&[2004]

, [Date].[Calendar].[Date])

, [Measures].[Internet Sales Amount], BDESC)

ON Rows

FROM [Adventure Works]

このクエリでは、集合関数 Descendants から返されたセットをパラメータとして別の集合関数 Order に渡す方法も示しています。

特定の条件に従ってセットにフィルタを適用することは、クエリを記述する場合に非常に役立ちます。このため、次の例に示すように、Filter (MDX) 関数を使用できます。

SELECT

[Measures].[Internet Sales Amount]

ON Columns,

//Returns a set containing all of the Dates beneath Calendar Year

//2004 in the Calendar hierarchy of the Date dimension

//where Internet Sales Amount is greater than $70000

FILTER(

DESCENDANTS(

[Date].[Calendar].[Calendar Year].&[2004]

, [Date].[Calendar].[Date])

, [Measures].[Internet Sales Amount]>70000)

ON Rows

FROM [Adventure Works]

また、他の方法でセットにフィルタを適用できるようにする、さらに高度な関数もあります。たとえば、次のクエリでは、TopCount (MDX) 関数によってセット内の上位 n 個の項目が返されることを示しています。

SELECT

[Measures].[Internet Sales Amount]

ON Columns,

//Returns a set containing the top 10 Dates beneath Calendar Year

//2004 in the Calendar hierarchy of the Date dimension by Internet Sales Amount

TOPCOUNT(

DESCENDANTS(

[Date].[Calendar].[Calendar Year].&[2004]

, [Date].[Calendar].[Date])

,10, [Measures].[Internet Sales Amount])

ON Rows

FROM [Adventure Works]

最後に、Intersect (MDX)Union (MDX)Except (MDX) などの関数を使用して、多数の論理的なセットの操作を実行することができます。次のクエリでは、後者 2 つの関数の例を示します。

SELECT

//Returns a set containing the Measures Internet Sales Amount, Internet Tax Amount and

//Internet Total Product Cost

UNION(

{[Measures].[Internet Sales Amount], [Measures].[Internet Tax Amount]}

, {[Measures].[Internet Total Product Cost]}

)

ON Columns,

//Returns a set containing all of the Dates beneath Calendar Year

//2004 in the Calendar hierarchy of the Date dimension

//except the January 1st 2004

EXCEPT(

DESCENDANTS(

[Date].[Calendar].[Calendar Year].&[2004]

, [Date].[Calendar].[Date])

,{[Date].[Calendar].[Date].&[915]})

ON Rows

FROM [Adventure Works]