PivotCube VCL的主要特點(diǎn)：

樹(shù)狀（層次）維度
它不但可以處理線(xiàn)性維度，還可以處理層次維度。我們把它們稱(chēng)為樹(shù)狀，是因?yàn)樗鼈兊臉?gòu)造就像標(biāo)準(zhǔn)的層次結(jié)構(gòu)—樹(shù)一樣，對(duì)于樹(shù)，所有Windows用戶(hù)應(yīng)該都很熟悉（目錄的結(jié)構(gòu)就像樹(shù)）。因此，你可以很容易地建立結(jié)構(gòu)類(lèi)似于Windows目錄的維度樹(shù)，它可以具有無(wú)數(shù)的葉子和節(jié)點(diǎn)。葉子和節(jié)點(diǎn)可以層層嵌套，最多可以有255層。
　
可擴(kuò)展的統(tǒng)計(jì)功能
要使用這個(gè)功能，你需要建立過(guò)度飽和立方體。但在此之前，你需要把PivotCube.ExtendedMode這個(gè)屬性設(shè)為T(mén)rue。如果一個(gè)立方體單元容納了很多事實(shí)表中的數(shù)據(jù)，那么統(tǒng)計(jì)計(jì)算就會(huì)變的很復(fù)雜。盡管如此，PivotCube仍然提供了基于沒(méi)有經(jīng)過(guò)簡(jiǎn)化處理的完整數(shù)據(jù)集的計(jì)算。這個(gè)特點(diǎn)使你可以計(jì)算關(guān)于當(dāng)前過(guò)濾集的準(zhǔn)確函數(shù)值。這個(gè)特點(diǎn)的獨(dú)特之處在于精確的計(jì)算而不是基于簡(jiǎn)化的統(tǒng)計(jì)公式的計(jì)算。因此，使用PivotCube，你總是可以得到準(zhǔn)確的計(jì)算結(jié)果。

PivotCube擴(kuò)展模式支持的統(tǒng)計(jì)功能如下：

• 最小值
• 最大值
• 中間值
• 四分之一
• 四分之三
• 四分點(diǎn)之和
• 四分點(diǎn)之差
• 四分點(diǎn)偏差系數(shù)
• 偏斜度
• 峰度
• 均方差
• 方差
• 偏差系數(shù)
• 平均誤差
• 平均絕對(duì)值差

    如果你不需要這個(gè)特殊功能，你可以使用PivotCube的標(biāo)準(zhǔn)模式來(lái)最小化立方體大小，內(nèi)存占用和提高速度。但在此之前，你需要把PivotCube.ExtendedMode屬性設(shè)為False。

PivotCube標(biāo)準(zhǔn)模式下的功能：

• 總和
• 計(jì)數(shù)
• 平均數(shù)

    可以很簡(jiǎn)單地從任何TDataSet的派生類(lèi)裝載數(shù)據(jù)
    使用TDataSet的派生類(lèi)作為數(shù)據(jù)源使你可以輕松地通過(guò)以下方式裝載數(shù)據(jù)：

• Borland數(shù)據(jù)庫(kù)引擎（BDE）
• ActiveX數(shù)據(jù)對(duì)象（ADO）
• Direct Oracle Access(DOA)
• IBObjects

    雖然你沒(méi)有必要使用“Group by”或者M(jìn)DX語(yǔ)句對(duì)數(shù)據(jù)進(jìn)行預(yù)處理，但是如果你希望從SQL-Server裝入PivotCube的數(shù)據(jù)記錄數(shù)最少，你就可以使用“Group by”語(yǔ)句

• 可以很容易地使用新數(shù)據(jù)更新已有的立方體（不需要重新構(gòu)造整個(gè)立方體）
• 就像在一個(gè)大型的OLAP服務(wù)器中，通過(guò)在已有的立方體中添加新的數(shù)據(jù)，而不是重新構(gòu)建整個(gè)立方體，你可以一步一步地構(gòu)造自己的立方體。對(duì)于那些需要處理經(jīng)常變化的數(shù)據(jù)的用戶(hù)來(lái)說(shuō)，這是非常重要的特色。

自定義的維包裝（dimension wrapping）

　　例如，“日期”可以被切分為季節(jié)、季度、白天/夜晚等等，或者，“地址”可以被切分為街道、郵政編碼、城市、村等。“姓” [比如smith]，“名” [比如John]和“部門(mén)”[比如managers]可以被綜合成單個(gè)字符串“Employer”[比如 John Smith Mgrs.]
　

可通過(guò)維度和度量值篩選

    OLAP最強(qiáng)大的功能之一是能幫助用戶(hù)進(jìn)行深刻和詳細(xì)地分析，從而得出正確的商業(yè)決策。PivotCube提供了強(qiáng)大的通過(guò)維度和度量值進(jìn)行篩選的功能。

• 通過(guò)維度進(jìn)行篩選以下面兩種方式進(jìn)行:
  
• 常規(guī)篩選（通常過(guò)濾掉不必要的數(shù)據(jù)）
• 增量式的篩選（只篩選出所選擇的一個(gè)必要的維度元素；但仍然可以分析無(wú)數(shù)的維度）。當(dāng)你只需要提供一個(gè)客戶(hù)的分析或者是一年的分析時(shí)，這種方式是非常方便的。
• 通過(guò)維度進(jìn)行篩選：對(duì)于活動(dòng)和不活動(dòng)的維來(lái)說(shuō)，都是可行的。例如，你不需要將一個(gè)維嵌入到一個(gè)活動(dòng)的片段中進(jìn)行分析。

PivotCube VCL key features

Tree-like (Hierarchical) dimensions

This feature allows working with not only linear dimensions, but with hierarchical dimensions as well. We call them “ Tree-like” because they are built like standard hierarchical structures – trees, good known to all Windows users (structure of directories is built like trees). So you can easily build dimensions which structure is branched like Windows directories and the quantity of nodes and leafs of the tree is unlimited. Both nodes and leafs can be embedded into nodes without any limitations up to 255 nested levels.

Extended set of statistical functions support.

To use this feature you need to build supersaturated cube. But before start building you need to set PivotCube.ExtendedMode to True. If one cube cell holds many values from the fact table - calculations of statistical functions get complicated. But nevertheless PivotCube provides calculations based on full data set without simplifying transformation of data. This allows calculating True function values regarding current filter sets. Unique thing of this feature is exact PivotCube calculations in contrary to calculations based on simplifying statistical formulas. So with PivotCube you always get True calculations as a result.

Statistical functions supported by PivotCube Extended Mode are as following:

• Min
• Max
• Median
• 1stQuartile
• 3rdQuartile
• InterQuartile
• Quartile Deviation
• Coeff of. Quartile Deviation
• Skewness
• Kurtosis
• Standard Deviation
• Variance
• Coeff. of Deviation
• Mean St. Error
• Mean Abs. Deviation

If you don’t need this special function you are able to minimize cube size, memory occupation and increase building speed using PivotCube Standard Mode. But before start building you need to set PivotCube.ExtendedMode to False.

List of aggregation functions allowed by PivotCube in Standard Mode:

• Summa
• Count
• Average

Simple load data from any TDataSet descendant.

Using TDataSet successors as datasource allows you easy load Data from:

• Borland database engine (BDE)
• ActiveX data objects (ADO)
• Direct Oracle Access (DOA)
• IBObjects
• Etc…

You don’t need to prepare you data with “Group by” or MDX clauses, but you may use “Group by” only if you wish to minimize loading data records into PivotCube from your SQL-server.
Easily upgrade saved cube with new data (without rebuilding total cube).
You can build your own cubes like in a 'large' OLAP servers – step-by-step, without rebuilding total cube, just adding new records into saved cube. It’s a very important feature especially for those users who work with often changing data.

Custom dimension wrapping

For example “Date” field can be splitted to Seasons, Quarters, Day/Night etc, or “Address” field can be splitted to street, zip-codes, city, village etc, fields “LastName” [e.g. Smith] “FirstName” [e.g. John] and “Department” [e.g. managers] can be combined to single string field “Employer” [e.g. John Smith mgrs.]

Filtering by dimensions and measures

• One of the most powerful OLAP features that helps to execute deep and detailed analysis to make business decisions is filtering. PivotCube provides powerful filtering by dimensions and filtering by measures.
  Filtering by dimensions is presented in 2 ways:
  • Custom filtering (ordinary filtering with the excluding of unnecessary data)
  • Incremental filtering (filtering with choosing one necessary dimension element only; quantity of analyzing dimensions is still unlimited). This is very convenient if you want for instance to provide analysis by one customer only or/and for only one year etc.
• Filtering by dimensions is available both in active and inactive dimensions. I.e. you don’t need to imbed a dimension into active slice to drill.