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. End-user can easy
build(define) own hierarchies for any dimension at runtime, save these hirarchies
into slice file and restore them when it needed. In that manner he can define
various hierarchies for same dimension. For examle, dimension Ware can be structured
by Ware-type or Ware-manufacturer etc.
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
Statistical functions supported by PivotCube Extended Mode are as following:
- Quartile Deviation
- Coeff of. Quartile Deviation
- Standard Deviation
- Coeff. of Deviation
- Mean St. Error
- Mean Abs. Deviation
- Custom measure calculation over full sorted data array.
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:
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)
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.
Large number of dimensions and measures. Now PivotCube
supports working with up to 255 dimensions and 1024 measures.
Saveand load built cubes into file or stream
(via IStream interface) for any use in future including publications.
Save and load built cubes into XML storage for
any use in future including publications in Web.
Save and load current cube slice into XML storage
for any use in future including publications in Web.
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
Simple sorting (with changing order) over any
measure and dimension in direct and back order
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.
Filtering by measures is an important thing for detailed analysis. PivotCube
provides a unique elaboration that allows excluding all data that don’t
correspond conditions of filtering out of aggregation. For example, you want
to exclude all sales less 10$ or to take into account only those customers who
bought only one piece of ware etc.
All filtering abilities can be used simultaneously and in any combination for
deep and detailed analysis.
Export PivotGrid into various media including
HTML,Excel,TStrinGrid,Windows metafile or Printer via own printing engine
Calculated measures in PivotMap at runtime available
via built-in formula language
Unicode characters support (require TNT unicode controls
and PivotCube VCL (VCL sources) package