Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02435861 2003-07-23
Static Drill-through Modelling
Field of the Invention
The present invention relates to a method of interactively searching a
database in such a manner that it is quick and easy. It is particularly
directed to
mechanisms that allow applications to present the user with summary
information. The present invention also relates to retrieving information from
a
database based on content aggregation, management and distribution.
Background of the Invention
A key to success in business today is to understand and effectively
manage the factors that drive an enterprise - a field known as Business
Intelligence (BI). Having critical information about such business drivers
allows
decisions that will significantly improve results. As organizations 'flatten',
that is,
the number of levels in their hierarchies decrease, these mission-critical
decisions are being made at lower levels - which means that almost every
employee in an enterprise needs quick, easy access to appropriate information.
In spite of this necessity, corporate information remains inaccessible for
many
employees - virtually locked away in data warehouses, data marts, enterprise
resource planning systems, and myriad corporate databases.
Early BI systems provided valuable information from these data stores by
retrieving transaction-level detail from On-line Tran;~action Processing
(OLTP)
databases. This capability relied on Information Technology (IT) specialists
building complex queries using languages like SQL. The advent of On-Line
Analytical Processing (OLAP), however, has given organizations more effective
and meaningful access to critical corporate data. Modern OLAP systems
consolidate and present summarized corporate information from a multitude of
sources. This technology allows users to view information in a business
context:
sales per quarter per sales representative, units shipped on time per city per
branch by air, and so on. By presenting corporate data so that it is related
to the
business structure, trends and anomalies can be more easily spotted and
addressed.
CA 02435861 2003-07-23
2
OLAP reports - interactive reports that are highly formatted, easily
deployed, and straightforward to use - deliver value to the entire
organization.
These reports accelerate the "Eureka moment" by exposing "sweet spots" of
information in a data set directly to decision makers, knowledge workers, and
information consumers. "Sweet spots" are selective pieces or collections of
information that provide decision makers with immediate critical insight into
business drivers. These OLAP reports can be regular status reports, but are
especially effective for key performance indicator (KPI) reporting, business
performance measurement reporting, and scorecard reporting, all of which are
becoming increasingly important to decision makers. A robust reporting
environment is required in which to perform these tasks.
Products such as PowerPIayT"" by ~ognos Inc. enable organizations to
create and deploy highly formatted, interactive ~LAIC reports. These reports
let
users easily measure, manage, and track improvements in business
l~ performance. They also allow easy distribution of this information across
the
enterprise. Decision makers throughout the organization now have the
information they need to significantly improve business results.
Because of the complexity of corporate organizations, and their data, it is
not unusual for a particular enterprise to deploy several different products
to
analyze their data. Often these products are from different vendors.
For companies that want to track performance and trends, or perform
scorecard-style management reporting (viz.: short, concise, and consistent),
there is an even more fundamental concern. !t can 'be extremely difficult to
understand "the big picture" when the only accessible reports focus on
transaction-level detail, because data in databases is organized for efficient
storage and administration, not for summary-level analysis or exploration. In
addition, data storage does not correspond to how the business is organized,
so
data must usually be manipulated and reformatted before the user can extract
useful information from it.
If, for example, managers want to explore company performance in terms
of product sales, a report that details the performance of individual sale
reps will
not help them spot overall trends. By reviewing summary information first,
such
as total sales per office or region, decision makers can more easily gain a
"big
CA 02435861 2003-07-23
3
picture" view of business performance. They can then drill down to lower-level
details to uncover what is driving these trends. Thus OLAP technology has
brought significant value to business decision making: OLAP systems store and
access data as dimensions that represent business factors like time, products,
geographical regions, and sales channels. This information is stored "multi-
dimensionally" - like a cube that can be viewed, turned, and explored from any
angle. The information is also presented in a business context, like 'number
of
customer complaints by product line in North America last quarter,' rather
than a
database context - so decision makers have immediate access to the information
they need to make the best decisions for the business.
Until recently, organizations have found it difficult to meet some of these
user requirements. However, with the advent of OLAP tools enterprise-wide
deployment of OLAP reports is now a reality. Cubes can be customized to
reflect
the dimensions and calculations (also called measures) based on data stored in
the original database most commonly used in a given organization. OLAP
reports are generated from one or more of these data cubes. Because each
cube contains a wide variety of dimensions and measures, a vast number of
reports can be built from the information in the cubes. The cubes can be
considered as master reports or a collection of components that can be
assembled to create a specific report.
With OLAP reports, the user's first view of the data is a 'top-level" one that
reveals patterns and trends at a glance. if users have identified issues in
this
summary-level information, OLAP reports enable them to fully explore and
analyze the data set from several perspectives or angles, to varying levels of
detail. The reports also enable users to 'slice and dice', drill down, drill
up, and
provide alternative graphical views of their data - something paper reports
cannot offer. (The term 'slice and dice' generally implies a systematic
reduction
of a body of data into smaller parts or views that will yield more
information. The
term is also used to mean the presentation of information in a variety of
different
and useful ways.)
OLAP reports that take an analyze-then-query approach allow decision
makers to access data the same way they identify and solve problems: by
CA 02435861 2003-07-23
4
reviewing totals or summary information first, then looking at the underlying
details by drilling down to transaction-level details whenever necessary.
There are two major stages in implementing a reporting solution. The first
step is to create OLAP cubes, the multidimensional structures that house
summary-level details of the corporate data. Typically, these cubes are
created
by IT specialists and deployed to information analysis and report authors. The
cubes are customized models of a business that reflect the unique
characteristics of the company. The structure of a cube is defined in terms of
dimensions and measures. Dimensions are hierarchical categories of information
like time, products, and geography. For example, the product dimension
hierarchy may be organized by product line, product group, and then by
individual product. Measures are the (results of) calculations based on the
original data that are used to track the business such as revenue, units sold,
and
cost of sales. In other words, measures are the numeric columns that present
the count or summation of particular values that users would like to see in
their
reports.
OLAP cubes generally contain only the dimensions and measures
relevant to a specific analysis. For example, sales analysis data and human
resources data would be housed in separate cubes. This ensures that cubes
remain manageable, not just in terms of their size but also in terms of the
clarity
of the information they contain. With appropriate tools, diverse but
compatible
cubes can be easily linked together so that users can move effortlessly from
one
cube to another, accessing information from all areas of the company.
Once OLAP cubes are created and deployedl, report authors have
everything they need to produce a wealth of OLAP ireports. The process for
authoring is extremely straightforward for all types cPf reports: status
reports that
reveal a snapshot of data; ad hoc reports that answer specific questions; and
business performance management reports that track KPIs (Key PerFormance
I ndicators).
Although OLAP reports can be distributed on paper, it is well known that
decision makers reap the most value when the reports are presented
electronically. There are typically three ways to explore data in an OLAP
report:
CA 02435861 2003-07-23
~ Drill downlDrill up: lJsers can explore a dimension hierarchically -moving
from summary-level information to the details and back - to gain fast answers
to
critical business questions. A financial manager who is concerned with rising
expenses will want to understand what parts of the company are particularly
problematic. By drilling down on a geographical dimension, they can move from
looking at expenses by country, by region, by office and then finally by
department. Drilling up is the reverse process, so th<~t the information
becomes
more summarized, and less detailed.
~ Drill-through and 'Slice and dice': Decision rnakers can interactively
la explore corporate data in any combination of dimensions, from many
different
angles or perspectives. For example, a sales manager can look at revenue
figures by product line, sales region, time period, or .'ales channel.
~ Graphical analysis: Users can choose from a variety of graphical
displays with which to depict the key factors that are driving the business
and
assist them in understanding the performance of various aspects of the
business.
In typical BI products the information defining the mapping of source and
target drill-through objects is spread over several report generating
applications,
sometimes obtained from different vendors, and uses data also provided by
2a different products, sometimes also from different verodors. 'This makes
administration difficult since changing drill-through behavior requires the
administrator to apply the changes to more than one report generating
application. It also means that it is very difficult to deploy or assess
dependencies for drill-through installations.
Another problem is that drill-through definitions are either saved in cubes
or saved in report definitions, so that when a change is made, such as a
report
being deleted or modified, i~t is very hard for the administrator to determine
what
cubes or other reports are dependent on the report tlhat is being deleted or
modified (i.e. are possible drill-through targets).
Further, drill-through objects are somewhat "closed in" or concealed, that
is, unintentionally, but effectively, hidden from external applications, so
that it is
difficult for third party applications to fully utilize them. The fact that
the data
describing a particular drill-through (the drill-through metadata) might be
CA 02435861 2003-07-23
6
distributed over several sources means that it is also difficult to integrate
it with
such third-party applications. Additionally, each of these applications likely
has
different data format requirements.
Summary of the Invention
The present invention, particularly when incorporated in a drill-through
modeling tool, solves or alleviates the above-mentioned problems, as well as
providing several other advantages as will be made clear in the following
description.
1~ According to one aspect of the invention, there is presented a method of
providing a drill-through service between two or more drill-through objects,
the
objects being drill-through sources and targets, the method comprising steps
of
defining one or more drill-through paths between drill-through objects, the
drill-
through path definitions being collected in a single structure; interfacing to
the
drill-through objects in a run-time environment using the collection of drill-
though
path definitions; and administering and maintaining the drill-through path
definitions, independently of applications using them.
Such a DMT also provides graphical displays of drill-through paths for a
DMT user or modeller. These displays show the parameter's and dependencies
of each drill-through path and allow users to obtain a quick overview of the
drill-
through network and further, they allow the tool user or modeller to confirm
drill-
through dependencies at a glance. Drill-through objects may thus be
manipulated and maintained in a graphical manner.
A further understanding of other aspects, features and advantages of the
present invention will be realized by reference to the following description,
appended claims, and accompanying drawings.
Brief Description of the Drawings
Preferred embodiments of the present invention will be described with
reference to the accompanying drawings, in which:
Figure 1 shows the computer and related systems and network within
which the invention may be practiced;
CA 02435861 2003-07-23
7
Figures 2a and 2b show two aspects of the relationships between the
data, the data modeling tool, the drill-through path rr~etadata, and the
report
generating applications; and
Figure 3 shows an example of drill-through nEawork in which the invention
may be used.
Detailed description of Preferred Embodiments
Typically, the invention is practiced within a network of computers and
related equipment, an example of which is illustrated in Figure 1. The data
are
stored in a data warehouse comprising files servers 100, 102, 104 and storage
devices 106, 108. The file servers 100, 102, 104 contain one or more software
report generating applications (not shown), includingi modeling and
administration tools using these data. The tile servers 100, 102, 104 are
accessed from workstations 110, 111, 112 over a network 120 which may be an
intranet, Internet, or any other arbitrary network topology. The workstations
110,
111, 112 may also contain application software (not shown).
In at least one preferred embodiment, a business modeling tool (BMT)
framework is enhanced with a drill-through path administration tool that
allows
the tool user or modeler to design a single drill-through solution comprising
all of
the metadata for drill-through sources and targets for an entire suite of
business
data analysis products, or in some cases a set of otherwise independent
business data analysis products. The drill-through metadata is that
information
necessary to describe all of the required or identified drill-through objects
and
their paths. ~ther embodiments permit access to this information (i.e. the
drill-
through metadata) to be extended to external applications as well. Yet further
embodiments permit the data accessed by the drill-tr~rough paths to be
resident
outside of the data warehouse. For example, historical commodity prices might
reside on a well-known Hypertext markup language (HTML) server attached to
and accessed via the Internet.
In describing the invention, it is helpful to understand how the BMT and
the drill-through service relate to the user of the data (who ultimately does
the
interpretation and analysis of that data) and to the report generating
applications
available, as well as to the data itself. The drill-through solution may be
CA 02435861 2003-07-23
considered as being implemented in two parts. The first part is the drill-
through
modeling tool that is used in performing the task of defining drill-through
paths
between various source and target objects. That is, metadata for a set of
drill-
through paths are defined and definitions of the drill-through paths are
aggregated at a single point and stored in a single structure or set of
structures..
The second part is the runtime environment that comprises those drill-through
services and queries used by report generating applications, i.e. interfacing
the
meta-data to various data collections, such as data cubes or data-based
reports,
which are derived from different report generating applications.
These parts are shown in Figure 2a where various report generating
applications, 211, 212, and 213 make use of a runtime environment 220 to
access metadata that defines valid drill-through path options using drill-
through
queries 225 on information stored as drill-through metadata 230. The drill-
through path metadata 230 is published through a s~:ries of interactions 235
by
the modeling environment 240. To reduce complexity of the Figure 2a the
common sources of data for the modeling and runtime environments are not
shown.
Figure 2b shows in more detail the mechanism for accessing the drill-
through path metadata 270 from a typical report generating application 250, by
making use of a drill-through service 260. When a drill-through path is
determined to be required by the application 250, under the control of a user
(not
shown) one or more queries through an application program interface API 255
are made of the drill-through service 260. The drill through service 2C0 reads
through a further API 265 the drill-through metadata 270 in the course of
responding to the one or more queries 255. Examples of queries that report
generating applications 250 (including those available from Business Objects
SA, Cognos Corporation, or Oracle Corporation) might issue through the API
are:
~ Generate a list of potential drill-through targets for a specified report
(source).
~ Generate a target report for a specified target and context information
(where an example of the context information is a row of data from the
report).
CA 02435861 2003-07-23
Figure 3 shows a representative example of a drill-through relationship
network that might be displayed in a diagram view within the BMT. Referring to
Figure 3, all drill-through sources and targets are represented as nodes. Any
node may be associated with any other node by me<~ns of a drill-through path,
shown diagrammatically as an arrow between the relevant nodes. This particular
example shows a number of reports generated by different report generating
applications. The type 1 reports Corporate Customer List 310, Customer Mailing
List 320, and Order Details 330, and the type 2 report Expenses 340 are
differentiated in that they are typically derived from separate report
generating
applications. In turn, these separate report generating applications may
typically
derive their data from different types of source (for example OLAP sources and
relational database sources). In addition, there are two cubes, Corporate
Sales
325, and Human Resources 335, derived from corporate data sources. One
'external' source of information is also shown, Current Stock Prices 315,
where
the reference is made via an Internet Uniform Resource Locator (URL). A
number of drill-through paths are identified, 312, 314, 316, 318, 322, 324,
326,
328 in which the arrowhead's designate the direction of the drill-through. A
list of
drill-through targets that is maintained within the drill-through service 250
of
Figure 2b is presented by the report generating application 250 of Figure 2b
either using the drill-through path names or the target names: The drill-
through
paths made available to the user by the report generating application are
identified by the tool user or by a modeller, or option<~Ily by an automatic
internal
process. If there is more than one path to a drill-through target then it is
the
designer's responsibility to provide meaningful drill-through path names to
assist
the tool user in choosing borrectly. Assigning these drill-through path names
is
optional. By way of example, paths 322, 326, 328 have names assigned to them
as follows: "By Employee" 322, "By Sales staff" 326 and "By Sales manager"
328. In cases where there are multiple drill-through paths to a single target
and
the drill-through paths do not have names assigned to them, the user of the
report generating application, when presented with a drill-through target list
dialogue, would see duplicate entries, one for each path.
It should be noted that in same cases a drill-through path might be
defined between two reports, such as 314, where the drill-through path extends
CA 02435861 2003-07-23
from one report, Corporate Customer list 310, to another report, Customer
Mailing list 320. In other cases, more than one drill-through path 326, 323 is
defined between two nodes, for example, between tlhe ~rder Details report 330
and the Expense report 340. There are no limits on connectivity of the drill-
S through paths, as long as they are logically correct - the selection of such
paths
is part of the drill-through designer°s task.
Adding drill-through source and targets is similar to an import action. It
will
be apparent that the details of the process of adding a drill-through object
(source or target) are specific to the type of object.
10 The following are the steps performed in one prefer,~ed embodiment to
add (or import) a drill-through object:
1 - locate and open the object.
2 - determine what parameters could be used as outputs and what
parameters could be part of the drill-through predicate (or search
condition).
3 - determine the parameters that could be uc;ed as inputs to the drill-
through object. For example, report columns, dimension levels, or
prompts.
4 - determine application-specific information, if any.
5 - create the drill-through object in the drill-through metadata structure.
Inputs to a report using drill-through objects can be anything that will act
as a valid filter to the report. The provision of the ability to drill through
from one
report to another, (or more generally from one docunnent to another), allows
the
user to switch rapidly from that first document/report to another relevant
documentlreport, while retaining some of the content: of the first
reportJdocument.
By way of example, we can consider a simple situation in which a report
has been defined that describes a product line of stoves (e.g. model#, colour,
price). The user wishes to open another report that describes sales
information
(date, totalSales), and to do so using the attribute 'colour' to relate the
two
reports. Even though none of the columns in the reports match, and the reports
therefore do not seem to be related, the definition of an appropriate join
(i.e
including colour) between the tables that are used in creating both reports
(by
mechanisms outside of this invention and discussion), means that colour can be
CA 02435861 2003-07-23
11
applied to the target report as a filter using the drill-through path. In this
case the
list of parameters used as an input to the drill-through path is: report
columns,
dimension levels, prompts and columns from other tables that have one or more
joins to tables referenced in the report (columns from associated tables).
It is in cases such as this one that the ability to define and maintain drill-
through paths provides significant benefit, since none of the reports involved
need be kept informed of changes to the relationships of the underlying data,
but
rather, only the drill-through paths need be kept up to date.. All of the
reports
dependent on the underlying data remain valid, resulting in a much reduced
maintenance burden.
In preferred embodiments of the invention the tool user is requested to
provide the location of the object that is to be imported (such as a report)
by
creating a drill-through path. One mechanism to achieve this is by the use of
a
graphical interface where the source and the target are associated with a line
on
a graphical representation of the system. Using such a mechanism, a dialog box
is then displayed which includes the source and target parameters in "list
boxes".
The tool user can "match up" the parameters by selecting equivalent (or
corresponding) pairs of parameters for the source and target from these list
boxes. The editing of the drill-through paths with the selection of
appropriate
parameters may optionally employ a form of dialog or wizard appropriate to the
user's competence. Optionally, the tool user may add parameter mapping
functions, where appropriate, using similar "list box" i:ools or their
equivalent.
In one preferred embodiment the drill-through modeling tool, after adding
(or importing) the object; automatically determines the possible drill-through
paths by examining the parameters of the imported object. It then attempts to
find parameters that match with parameters of other drill-through objects.
This
list of potential drill-through paths is then presented to the tool user or
modeller
in a manner allowing the selection of one or more desired drill-through paths.
In further preferred embodiments the adding process includes a feature to
attempt to determine the physical source on which a parameter is based. This
is
carried out automatically and the resultant matches are presented as a list.
The
tool user then selects or deselects the matches from those presented. In these
embodiments, if the source and target parameter names do not match exactly
CA 02435861 2003-07-23
12
then other information regarding the parameters is used as a clue as to which
source and target parameters match. For example a parameter may represent a
report column where that column is based on a database column. The reference
to that database column is saved and used as a clue or'hint' to assist the
tool
user when trying to establish a mapping from a parameter from one drill-object
source to target. As before, the tool user selects or deselects from a list to
make
a selection.
Typically, drill-through objects (source and targets) have their
implementation-specific actions encapsulated within a dynamically shared
library, within what are known as plug-ins. Alternatively, the objects and
their
actions may form a module in a program library. Each plug-in or module is
responsible for at Least the following functions:
- Importing (or adding), which is the act of determining the parameters
required to define the drill-through object.
- Providing the actions and logic that the drill-through service needs to send
either to perform the drill-through action or to return the required data to
the
calling client of the drill-through service. When the drill-through service is
queried and asked for a list of targets based on a drill-through predicate (or
search condition) and its source, the drill-through server responds with a
list
of targets plus other associated information (such as a URL). The inclusion of
a URL means that the client making the query could, based on the response,
'launch' the URL, thereby accessing the data contained in the referred-to
location. The drill-through service also has actions that can be invoked by
the
client when it is required to perform the action on behalf of the client.
Typically, the drill-through service invokes the URL (for example by sending
back a redirect HTML page to the client's browseir). The plug-in specific to
the
drill-through target determines which action is performed. In some
embodiments, the action queries one or more third party databases or
applications (or more strictly the data or reports associated with or
generated
by the applications).
- Validating (when invoked by the modeling application), the existence of
the target and confirming that the parameter and other properties still
applylexist.
CA 02435861 2003-07-23
13
- Optionally translatiing andlor reformatting of data between any two
compatible applications.
The use of a plug-in, or an equivalent program library, means the
metadata for each drill-through path is stored in a single place or structure,
in a
fixed and defined format, thereby simplifying the updating ~f such drill-
through
paths, and further simplifying the disclosure and publication of APIs and
their
related drill-through path metadata for the use of third-party applications if
necessary.
While various preferred embodiments of the present invention have been
described above, it should be understood that they have been presented by way
of example only, and not limitation. It will be understood b~,y those skilled
in the
art that various changes in form and details may be made therein without
departing from the spirit and scope of the invention as defined in the
appended
claims. Thus, the breadth and scope of the present invention should not be
IS limited by any of the above-described exemplary ennbodiments, but should be
defined only in accordance with the following claims and their equivalents.
