Systems and methods for general aggregation of characteristics and key figures

Abstract

Computer-implemented methods, computer systems, and computer programs product are provided for automated generic and parallel aggregation of characteristics and key figures of unsorted mass data being of specific economic interest, particularly associated with financial institutions, and with financial affairs in banking practice. The parallel aggregation may reduce the amount of data for a customer defined granularity for the purpose of facilitating the handling of raw data related to all areas of credit risk management in banking practice. Moreover, the computing power of software and the software performance run time, respectively, may be improved in the case of mass data.

Description

I. TECHNICAL FIELD

The present invention generally relates to electronic data processing. More particularly, the invention relates to computer-implemented methods, computer systems and computer program products for automated generic and/or parallel aggregation of characteristics and key figures of mass data associated with, for example, financial institutions and with financial affairs in banking practice.

II. BACKGROUND

As international financial markets expand, global concerns over the soundness of banking practices are driving stringent new requirements for bank-level management, regulatory control, and market disclosure.

Data processing systems in banking are provided with software tools, for example, SAP proprietary software tool solutions in banking such as the SAP solution for the new Basel Capital Accord (Basel II) that builds on the proven capabilities of the SAP for Banking solution portfolio, that enable financial entities and other users to pursue these requirements.

The SAP solution for the new Basel Capital Accord (Basel II) represents a risk-sensitive framework that provides capabilities for calculating risk exposure and capital, for managing market risk, interest risk, or liquidity risk, and for calculating and managing all areas of credit risk, helping to facilitate the handling of mass data, particularly being of specific economic interest and associated with financial institutions and with financial affairs in banking practice.

Moreover, software tool solutions for banking systems including capabilities for computing descriptive statistics are needed to efficiently analyze large amounts of given data (mass data) while managing large and complex projects. Within that scope, mass data are often required to be aggregated according to a customer defined granularity. Accordingly, aggregations can be computed for characteristics (lexicographic min, max) and key figures (min, max, count, sum, avg, variance, std, var %) using prior art software tool solutions.

However, there still remains the need to improve the computing power of software and software performance (i.e., run time performance), respectively and, in particular, when it comes to large amounts of data (mass data) to be aggregated effectively that can not be handled in the main memory of a data processor.

SUMMARY

Embodiments consistent with the present invention meet the above-identified needs by providing, for example, computer-implemented methods for automated generic and parallel aggregation of characteristics and key figures of mass data, particularly associated with banking practice, that can be easily integrated into existing credit risk platforms such as, for example, the above mentioned SAP solution for Basel II.

Embodiments of the present invention also relate to computer systems and computer program products for automated generic and parallel aggregation of characteristics and key figures of mass data. Further, embodiments of the invention relate to data carriers readable by a computer, the data carrier storing a plurality of instructions implemented by a computer program for causing the processing means of a computer system to execute computer-implemented methods and/or other embodiments of the present invention.

Moreover, an embodiment of the present invention provides a computer-implemented method for automated generic and parallel aggregation of characteristics and key figures of mass data associated with banking practice, that are not assumed to be a priori sorted in respect to a free selectable granularity before applying the computer-implemented method.

An embodiment of the present invention also provides a computer-implemented method that can optionally perform the automated generic aggregation of data either in linear or in parallel processing mode, thereby noticeably improving the computing power of software, as preferably in the case of mass data, depending on the capacity utilization of a data processing system.

In accordance with the invention as embodied and broadly described herein, a computer-implemented method may be provided for automated generic and parallel aggregation of characteristics and key figures of mass data whose structure is unknown, particularly associated with financial institutions and with financial affairs in banking practice, provided by different databases of different data sources. The inventive method may reduce the amount of data to a customer defined granularity by computing aggregations on key figures within the scope of an iterative process, repeatedly processing a parallel aggregation algorithm including parallel processing steps for merging, reorganizing, sorting and aggregating data records.

In one embodiment of the present invention, the aggregation may be computed on predetermined key figures using predetermined aggregation operations selected from a function pool and/or customer defined aggregation operations may be defined by input means using the predetermined aggregation operations.

In another embodiment of the invention, the aggregation is computed on costumer defined key figures to be defined by input means using the predetermined aggregation operations selected from a function pool and using the predetermined aggregation operations and/or said customer defined aggregation operations.

In yet another embodiment of the invention, the aggregation algorithm can run in parallel processing mode for mass data, thereby noticeably improving the computing power of software, but if required, depending on the capacity utilization of a data processing system, the processing of the aggregation algorithm can optionally run in linear processing mode.

According to another embodiment of the invention, the aggregation algorithm of the present invention can easily be integrated into other processes, e.g., as a pre-processing before a data extraction of business area information to a business information warehouse of a company, thereby separating the results of already aggregated mass data for the purpose of visualizing data of specific economic interest.

According to still another embodiment of the present invention, the aggregation algorithm can be applied to prior art software solutions in the context of an ad hoc reporting for descriptive statistics.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate several embodiments of the invention and together with the description, serve to explain the principles of the invention. In the drawings:

FIG. 1 illustrates a schematic view of an exemplary computer-implemented method for automated generic and parallel aggregation of characteristics and key figures of unsorted mass data;

FIG. 2 illustrates a simplified flow chart of an exemplary computer-implemented method showing the steps for automated generic and parallel aggregation of characteristics and key figures of unsorted mass data;

FIG. 3 illustrates a flow chart of an exemplary method showing the steps for the aggregation of records within a single data package;

FIG. 4 a illustrates an example of use for raw data, showing a work list of M=12 data records associated with financial institutions and with financial affairs in banking practice;

FIG. 4 b illustrates an example of granularity characteristics/granularity levels i of granularity characteristics;

FIG. 5 illustrates an example of use for the parallel aggregation algorithm illustrated in FIG. 2, wherein the original amount of data records shown in FIG. 4a is reduced to N=4<M=12 data records for a customer defined granularity as it is set out in FIG. 5 referring to “search result”;

FIG. 6 illustrates an example of use for the parallel aggregation algorithm illustrated in FIG. 2, wherein the original amount of data records shown in FIG. 4a is reduced to N=4<M=12 data records for a customer defined granularity as it is set out in FIG. 5 referring to “search result”, and wherein another compromise of performance is made compared to the preceding example of use of FIG. 5;

FIGS. 7A-7C illustrate an example of a work list, consistent with an embodiment of the present invention; and

FIGS. 8A-8C illustrate another example of a work list, consistent with an embodiment of the present invention.

DETAILED DESCRIPTION

Reference will now be made in detail to the present invention, examples of which are illustrated in the accompanying drawings in which like reference numbers refer to corresponding elements.

The present invention does not only relate to computer-implemented methods for automated generic and parallel aggregation of characteristics and key figures of unsorted mass data associated with financial institutions and with financial affairs, but also to data processing systems, computer program products that can be stored on a computer readable data carrier, and data carriers.

Data processing systems (computer systems) consistent with embodiments of the invention may comprise a single data processor or a plurality of data processors via inter-computer network, each data processor including processing means (processor), storage means (memory), bus means (bus), network means (network), interface means, input means and output means (input and output devices). The computer system may also be simply a server.

The data processor is, for example, a conventional desktop computer, a multiprocessor computer, or the like. The processor is, for example, a Central Processing Unit (CPU), a Micro Controller Unit (MCU), Digital Signal Processor (DSP), or the like.

Storage means are in particular provided for storing the specified mass data. Storage means symbolizes any memory means for temporarily or permanently storing data and instructions. Although memory is conveniently illustrated as part of computer, memory function may also be implemented in network, in computers and in processor itself, e.g. cache, register, or elsewhere. Memory can be, for example, a Read Only Memory (ROM), Random Access Memory (RAM), or a memory with other access options. Memory is physically implemented by computer-readable media, for example: (a) magnetic media, such as hard disk, floppy disk or other magnetic disk, tape or cassette tape; (b) optical media, such as optical disk (CD-ROM, DVD); (c) semiconductor media, like DRAM, SRAM, EPROM, EEPROM, or the like.

Memory means may further store support modules, for example, a Basic Input Output System (BIOS), an Operating system (OS), a program library, a compiler or interpreter, and a text processing tool.

Input means symbolizes any device for providing data and instructions for processing by computer, for example, a keyboard or pointing device such as a mouse, trackball or cursor direction key.

Output means symbolizes any device for presenting results of aggregated data packages, for example, a monitor or a display, for example, a Cathode Ray Tube (CRT), Flat Panel Display, Liquid Crystal Display (LCD), or printer.

Bus and network provide logical and physical connections by conveying data and instruction signals. While connections inside computer are conveniently referred to as “bus”, connections between computers are referred to as “inter-computer network”. Optionally, network comprises gateways being devices (computers) that specialize in data transmission and protocol conversion, allowing users working in one network to access another network.

Networking environments (as network) are commonplace in offices, enterprise-wide computer networks, intranets and the internet (i.e. world wide web). Network can be a wired or wireless network. To name a few network implementations, network is, for example, a local area network (LAN), a wide area network (WAN), a public switched telephone network (PSTN), an Integrated Services Network (ISDN), an infra-red (IR) link, a radio link, like Universal Mobile Tele-communications System (UMTS), Global System for Mobile Communication (GSM), Code Division Multiple Access (CDMA), or satellite link.

Transmission protocols and data formats are known as, for example, transmission control protocol/internet protocol (TCP/IP), hyper text transfer protocol (HTTP), secure HTTP, wireless application protocol, unique resource locator (URL), unique resource identifier (URI), hyper text markup language HTML, extensible markup language XML, extensible hyper text markup language XHTML, wireless application markup language (WML), etc.

Interface means (interfaces) for linking together the data processing units of a data processing system are well known in the art. An interface can be, for example, a serial port interface, a parallel port interface, a universal serial bus (USB) interface, an internal or external modem.

Computer program products consistent with embodiments of the invention may comprise a plurality of instructions for causing the processing means of a computer system to execute the method steps of the invention specified hereinafter with more detail. In other words, computer program product defines the operation of computer and its interaction in inter-computer network. For example, a computer program product may be available as source code in any programming language, and as object code (binary code) in a compiled form. Persons skilled in the art can use computer program products in connection with any of support modules (e.g. compiler, interpreter, operating system). The computer program products may be stored in memory hereinafter referred to as data carrier.

For the communication between computer program products and computers, data carriers may be conveniently inserted into an input device. A data carrier may be implemented as any computer readable medium. Generally, a data carrier is an article of manufacture comprising a computer readable medium having readable program code means embodied therein for executing the method steps of the present invention. Furthermore, program signals can also embody computer programs. Program signals may be transmitted via inter-computer network(s) to a data processor.

FIG. 1 illustrates a schematic view of an exemplary computer-implemented method, consistent with an embodiment of the invention. The exemplary method may be used for automated generic and parallel aggregation of characteristics and key figures of unsorted mass data, in particular being of specific economic interest and associated with financial institutions and with financial affairs in banking practice. The mass data (“input data”) whose structure is unknown include a plurality of M data records, wherein M represents a large amount of data records to be aggregated that can not be handled in the main memory of a data processor. The mass data (“input data”) may further comprise packetized blocks of data provided by different databases of different accessible data sources, including sets of rows and sets of columns, each row corresponding to a record, and the columns including fields of predetermined granularity characteristics and fields of predetermined key figures. Generally speaking, the generic aggregation of characteristics and key figures aims at the reduction of said mass data according to a given customized granularity. Due to the plurality of M data records, said mass data are customized as packages including M_p<M data records as it is illustrated in the upper block of FIG. 1 referred to as “Built packages” before being assigned to the parallel aggregation algorithm. The built data packages (package 1, package 2 . . . package n) are assigned to different jobs so that each job includes a plurality of data packages. A job or a plurality of jobs can be processed in a parallel processing mode, thereby noticeably improving the computing power and run time performance of software, respectively, either using a single data processor or a network of data processors by applying the method steps illustrated in the lower block of FIG. 1. But if required, depending on the capacity utilization of a data processing system, the processing of said aggregation algorithm can optionally run in linear processing mode, thereby aggregating and merging packages within a job sequentially. The method steps of the aggregation algorithm illustrated in the lower block of FIG. 1 are explained in detail below.

FIG. 2 illustrates a simplified flow chart of an exemplary computer-implemented method showing the steps for automated generic and parallel aggregation of characteristics and key figures of unsorted mass data.

In step 10, the computer-implemented method begins with a selection of investigated mass data (“input data”) including said plurality of M data records to be aggregated, said mass data being provided by different accessible primary databases of different accessible data sources. Having finished the selection of mass data due to selection criteria, the variously selected blocks of packetized mass data are assigned among each other and the result of assignment is stored to a global database.

Thereupon, some customizing of the selected mass data is required for defining granularity characteristics and aggregation operations to be carried out by the processing means of a data processing system for computing fields of key figures.

Moreover, the selected mass data are prepared as data packages according to a customer defined package size including M_p<M data records in a pre-processing step before reading said mass data into the processing means of a data processing system.

In step 20, the packaged data can be additionally enriched in a parallel pre-processing with data from an accessible single secondary database or from accessible secondary databases, subsequently saving the results of enrichment to those local databases of the respective data processors where the data are to be processed.

Subsequent to the packaging, the data packages are read into the data processing means of a data processing system to be processed within jobs, each of the jobs including a plurality of data packages.

A job or a plurality of jobs can be processed in a parallel processing mode either using a single data processor or a network of data processors.

Up to this point, there is not identified one granularity level i corresponding to i=0. By assigning the data packages to the parallel aggregation algorithm 30, a first parallel processing begins with step 40, wherein at first said customized granularity characteristics are identified so as to obtain levels i of granularity characteristics. Having identified said granularity characteristics within said data packages, thereby accomplishing the criterion i>0, the records of each data package are sorted for a given order of said granularity characteristics, and subsequently aggregated for said key figures by using customized aggregation operations, thereby reducing the amount of records in said data packages to N_a<M_p (smaller than the maximum size M_p). Following the aggregation, the results of each aggregated data package are saved to those local databases of the respective data processors where the data are processed. Thereupon, the aggregated packages are split into several smaller sub packages including N_sp data records and the size (number of records) and the first and the last record of each sub data package is stored to a global result database. Hereafter, the identification of adjacent packages based on these small sub data packages is executed by checking the termination criterion for the loop i=i+1 (“not in parallel”) being:

• • if key_pos1,xε(key_pos1,y;key_{pos max:y}) then continue else terminate,

wherein pos1 illustrates the first position of a data package, posmax illustrates the last position of a data package, and x, y illustrates the number of a data package, thereby comparing the key of the first record of each data package with the first and the last record of all the rest of data packages (thus comparing all combinations x, y). If said criterion for terminating the loop i=i+1 is not accomplished, meaning that the conditional inquiry is true, the data packages are assigned for rebuilding new data packages.

The underlying idea of splitting aggregated data packages into sub data packages is to improve the expressiveness of key information, and thereby to improve the identification of adjacent data packages based on their respective key information. Since only the data package size and the key information of the first and the last record of each data package are stored to a global database while all other data records are not considered, the following interests working in opposite directions must be kept in mind. While large package sizes are ideal for aggregation, the key information of the first and the last record of each large data package is not representative for all the rest of data records within said data package. On the other hand, if the data packages are very small, then the first and last record of each data package is more or less representative for all the rest of data records. But by reducing package sizes, the efficiency of aggregation diminishes due to the fact that there is not much to aggregate in small data packages.

Thus, the point is to meet the above-identified two interests working in opposite direction by approaching an efficient compromise of performance allowing to aggregate relatively large data package sizes, and subsequently split the aggregated data packages into smaller sub packages for the purpose of identifying adjacent sub data packages.

The relation of the maximum data package size M_p and the size of sub data packages N_sp depends on the degree of fragmentation and the degree of aggregation of the unsorted input data.

The effect of this approach of splitting aggregated data packages into sub data packages becomes the more important the less sorted the input data are, and the lower the degree of aggregation is, or in other words, the lower the reduction of the number of data records is.

In step 50, the aggregated packages may be assigned to a second parallel processing of the aggregation algorithm 30 for merging adjacent packages, thereby rebuilding new data packages, wherein adjacent packages are those packages with keys of the first record which are closest together. By merging these small data packages the maximum allowed package size M_p is restored again. Hereupon, the new data packages (merged packages) are stored to local databases of the respective data processors where the data are processed. Subsequently, the new data packages are assigned again to the above mentioned first parallel processing for reorganizing and sorting, and thereafter aggregating said new data packages for key figures by using said customized aggregation operations.

After each loop cycle, the conditional inquiry for terminating the loop i=i+1 is checked for all combinations (x, y) anew, repeatedly executing the loop i=i+1 while the termination criterion is not accomplished, meaning that the conditional inquiry is true, else, after accomplishing said criterion, i.e., all the data packages are disjoint with regard to the granularity characteristics, terminating the loop.

Finally, the packaged data can be additionally enriched in a parallel post-processing step 60 with data from an accessible secondary database or from accessible secondary databases, subsequently saving the results of data packages to a global result database.

FIG. 3 illustrates a flow chart of an exemplary method, consistent with an embodiment of the invention, showing the steps for the aggregation of records within a single data package after entering the aggregation algorithm 30 of FIG. 2, for example. At first, in step 70, there is not identified one level of granularity characteristics, which is symbolized by i=0. Furthermore, before aggregating for the first time within the scope of a first iteration, the summary table referred to as itab in which the aggregation result is stored to is empty. At this point, the records within the data package are assigned to a first parallel processing, wherein the fields of granularity characteristics are identified according to a customer defined granularity so as to obtain levels i (i=1 . . . n) of identified granularity characteristics. Having identified said granularity characteristics, thereby accomplishing the criterion i>0, the records of said data package are sorted for a given order of said granularity characteristics as, for example, illustrated in FIG. 5 referring to “search result”.

By entering the loop 85, the records are assigned to the approach for sequentially aggregating the unique granularity levels i using predetermined and customized aggregation operations. Beginning with the first granularity level i=1 in step 80, the level i=1 is compared with the maximum level n inquiring the condition being (i>n ?) in step 90. As long as the condition (i>n ?) is not accomplished, meaning that the conditional inquiry is false, and thus i having a value less than or equal n, in method step 100, the records of the data package tab(i) corresponding to the appropriate granularity level i=1 are aggregated for a specific key figure Xj by using predetermined aggregation operations (operator j), thereby entering an internal loop 95.

Subsequently, in step 110, the aggregated key figure Xj^| is moved to the structure str1. Thereafter, in step 120, it is inquired if the aggregation of data records for key figures in respect to the appropriate granularity level i=1 is completed. If the conditional inquiry is not accomplished, the records of the data package corresponding to level i=1 are assigned again to a subsequent aggregation in respect to another key figure using another operation, repeatedly executing this approach of aggregation steps until all selected aggregation operations are conducted, else, leaving the internal loop 95.

In step 130, customer defined aggregation operations can be applied using, for example, an SAP-BAdl aggregation technique (commercially available form SAP AG Walldorf, Germany), subsequently saving the results to the structure str1, wherein previous results may be changed. Thereupon, having completed the aggregation algorithm for the appropriate level i=1, the structure str1 is appended to the summary table itab. This approach for executing the loop 85 is to be applied to all remaining granularity levels i up to and including the maximum number of i (i=2, 3, 4 . . . n).

By accomplishing said criterion in step 90 for leaving the loop 85 being (i>n ?), in step 150, a global administration table is filled with itab-information. Finally, in step 160, the summary table referred to as itab is saved to a local database.

FIG. 4 a illustrates an example of use for raw data showing a work list including M=12 records associated with, for instance, financial institutions and with financial affairs in banking practice to be applied to the parallel aggregation algorithm 30 of FIG. 2, for example. The work list includes sets of rows and sets of columns, each row corresponding to a record, and the columns including fields of predetermined granularity characteristics, and fields of predetermined key figures.

Furthermore, the records are sorted according to a given order of granularity characteristics as set out in FIG. 4b under “granularity characteristics/“granularity levels i of granularity characteristics”.

FIG. 5 illustrates an example of use for the aggregation using a processing tool based on the parallel aggregation algorithm 30 of FIG. 2, for example. The aggregation of the raw data illustrated in FIG. 4a including M=12 data records reduces the amount of data to 4<M=12 data records according to the customer defined granularity, as it is set out in FIG. 5 referring to “search result”.

The granularity fields including granularity characteristics are characterized by “rating method” and “rating segment”. The fields of key figures are characterized by the columns “financial statement key figure 1” and “financial statement key figure 2”.

The data package size is determined through customizing. Contrary to the preceding statement that large data package sizes are ideal for aggregation, whereas small data package sizes are ideal for reorganization, in this example of use only one single package size can be determined, meaning that the data package size M_p for aggregation is identical to the sub data package size N_sp for reorganization. Therefore, in this example of use a less efficient compromise of performance has to be chosen to meet said opposite demands. The customized package size is determined by M_p=4 corresponding to the maximum number of granularity levels i, as it is shown in the table of FIG. 5 referring to “search result.”

In step 200 (see, e.g., Table 2 below and FIG. 7A), the raw data shown in the original work list of FIG. 4a are exemplarily arranged by the key figures in the column “financial statement key figure 1” in ascending order so as to demonstrate a work list of unsorted records to begin with. Due to the customized data package size of M_p=4, the M=12 data records of said work list are split in three data packages, in data package 1, data package 2 and data package 3, each data package as a result having 4 data records.

Furthermore, for the exemplification of the parallel aggregation algorithm as illustrated in the example of FIG. 2 on the basis of the concrete example and to simplify matters, only the granularity fields characterized by “rating method” and “rating segment”, the fields of key figures characterized by the columns “financial statement key figure 1” and “financial statement key figure 2”, and the field currency are taken into consideration. All the rest of fields remain empty. Hereinafter, Table 2 illustrates the outcome of this reorganization and simplification of said original work list shown in FIG. 4a.

TABLE 2
STEP: 200
		financial	financial
		statement key	statement key
rating method	rating-segment			currency	package
insurances	life insurances	1620	865860	EUR	1
credit institutions	Landesbanken (form of banks)	1912	809485	EUR	1
credit institutions	Sparkassen (form of banks)	2860	456825	EUR	1
credit institutions	Sparkassen (form of banks)	3254	693677	EUR	1
insurances	casualty insurances	3346	729541	EUR	2
credit institutions	Landesbanken (form of banks)	3393	542616	EUR	2
insurances	life insurances	5966	670365	EUR	2
credit institutions	Landesbanken (form of banks)	6135	166310	EUR	2
credit institutions	Sparkassen (form of banks)	8149	484449	EUR	3
insurances casualty	insurances	8683	824001	EUR	3
insurances life	insurances	8715	247374	EUR	3
insurances	casualty insurances	8916	35040	EUR	3

In step 210 (see, e.g., FIG. 7A), the data packages are assigned to the parallel processing of the aggregation algorithm 30 of FIG. 2. Within the scope of a first iteration (Iteration Nr. 1), the parallel processing begins with the method step 40 of FIG. 2. Up to this point, there is not identified one granularity level i mentioned above, which is symbolized by i=0 in FIG. 2. Therefore, at first, the fields of granularity characteristics labeled “rating method” and “rating segment” are identified so as to obtain levels i of granularity characteristics within said data packages, thereby accomplishing the criterion i>0. The maximum reachable number of granularity levels i per data package is i=4 due to the table of FIG. 5 mentioned above. By sequentially comparing the above mentioned granularity characteristics shown in the table of FIG. 5 with the data records of each of the three data packages, thereby beginning with the first row of granularity characteristics of the table of FIG. 5 characterized through “credit institutions/private banks” and ending with the forth row characterized through “insurances/casualty insurances”, in the example of use there appear in total three matches in each of the three data packages, what as a result corresponds to a granularity level of i=3 of identified granularity characteristics for each data package.

Subsequently, the data records within all of the three data packages are sorted according to the given order as set out in the table of FIG. 5. The outcome of this sorting is illustrated below in Table 3.

TABLE 3
STEP: 200
		Sum	Minimum
		financial	financial
		statement key	statement key
rating method	rating-segment			currency	package
credit institutions	private banks	1912	809485	EUR	1
credit institutions	public banks	3254	693677	EUR	1
credit institutions	public banks	2860	456825	EUR	1
insurances	life insurances	1620	865860	EUR	1
credit institutions	private banks	6135	166310	EUR	2
credit institutions	private banks	3393	542616	EUR	2
insurances	life insurances	5966	670365	EUR	2
insurances	casualty insurances	3346	729541	EUR	2
credit institutions	public banks	8149	484449	EUR	3
insurances life	insurances	8715	247374	EUR	3
insurances casualty	insurances	8683	824001	EUR	3
insurances casualty	insurances	8916	35040	EUR	3

As illustrated in Table 3, the first two rows of data package 1 and data package 2 appear to have identical granularity characteristics. In data package 3, the last two rows include identical granularity characteristics.

Thereafter, these rows are aggregated for the key figures (Xj) characterized through “financial statement key figure 1” and “financial statement key figure 2” by applying appropriate aggregation operations (operators j) to the respective key figures, wherein said aggregation operations being predetermined or customized aggregation operations. In this case the matches are added up in respect to key figure 1, and in respect to key figure 2 the minimum value is taken over. All three data packages are processed simultaneously due to the parallel processing. As a result, the number of data records within all of the three data packages is reduced to N=3<M_p=4 data records, which is illustrated below in Table 4.

TABLE 4
STEP: 210
		Sum	Minimum
		financial	financial
		statement key	statement key
rating method	rating-segment			currency	package
credit institutions	private banks	1912	809485	EUR	1
credit institutions	public banks	6114	456825	EUR	1
insurances	life insurances	1620	865860	EUR	1
credit institutions	private banks	9528	166310	EUR	2
insurances	life insurances	5966	670365	EUR	2
insurances	casualty insurances	3346	729541	EUR	2
credit institutions	public banks	8149 484449	EUR	3
insurances	life insurances	8715	247374	EUR	3
insurances	casualty insurances	17599	35040	EUR	3

Thereupon, in step 220 (see, e.g., FIG. 7B), after saving the results of each data package in a database, the identification of adjacent data packages is conducted by checking the termination criterion of the loop i=i+1 (“not in parallel”) being:

• • if key_pos1,xε(key_pos1,y;key_{pos max:y}) then continue else terminate,

wherein pos1 illustrates the first position of a data package, posmax illustrates the last position of a data package, and x, y illustrates the number of a data package, thereby comparing the key of the first record of each data package with the key of the first and the last record of all the rest of data packages (thus comparing all combinations x, y). If said criterion is not accomplished, meaning that the conditional inquiry is true, the data packages are assigned for rebuilding new data packages.

Beginning with the key of the first record of data package 1, the comparison of data package 1 and data package 2 results in that the key of the first record of data package 1 is equal to the key of the first record of data package 2. As a result, interpreting the conditional inquiry for the loop, the key of the first record of data package 1 is an element of the amount of data in data package 2, or furthermore interpreted, data package 1 and data package 2 intersect, and thus they are identified as adjacent packages. Consequently, as the termination criterion for the loop is not accomplished, data package 1 and data package 2 are assigned for rebuilding a new data package 1. Since data package 1 and data package 2, respectively, include 3 records, the data package size of the new data package 1 including M_m=6 records exceeds the maximum package size of M_p=4, which is acceptable. The data package 3 remains unmodified.

In step 230 (see, e.g., FIG. 7B), the aggregated data packages are assigned to the second parallel processing of the aggregation algorithm 30 of FIG. 2 illustrated by the step 50 within the scope of a second iteration (Iteration Nr. 2) for merging the adjacent data packages of step 210. Having merged said data package 1 and data package 2 to a new data package 1, the data records of the remaining two data packages are assigned again to the above mentioned first parallel process illustrated by step 40 of FIG. 2 within the scope of a second iteration (Iteration Nr. 2), wherein the data records of the remaining two data packages are reorganized in parallel processing mode, and thereafter sorted again according to the given order for said granularity characteristics as illustrated in the table of FIG. 5 and in FIG. 4b, respectively. This outcome of this reorganization and sorting is illustrated hereinafter in Table 5.

TABLE 5
STEP: 230
		Sum	Minimum
		financial	financial
		statement key	statement key		old	new
rating method	rating-segment			currency	package	package
credit institutions	private banks	1912	809485	EUR	1	1
credit institutions	private banks	9528	166310	EUR	2	1
credit institutions	public banks	6114	456825	EUR	1	1
insurances	life insurances	1620	865860	EUR	1	1
insurances	life insurances	5966	670365	EUR	2	1
insurances	casualty insurances	3346	729541	EUR	2	1
credit institutions	public banks	8149	484449	EUR	3	2
insurances	life insurances	8715	247374	EUR	3	2
insurances	casualty insurances	17599	35040	EUR	3	2

Thereupon, the aggregation for the key figures using the predetermined aggregation operations is conducted anew, wherein as a result, the size of the new data package 1 decreases from M_m=6 to N_a=4 according to the customer defined granularity as illustrated in FIG. 5 referring to “search result”. Following the aggregation, the results of the remaining data packages are saved in a database. The outcome of this aggregation is illustrated hereinafter in Table 6.

TABLE 6
STEP: 230
		financial	financial
		statement key	statement key
rating method	rating-segment			currency	package
credit institutions	private banks	11400	166310	EUR	1
credit institutions	public banks	6114	456825	EUR	1
insurances	life insurances	7586	670365	EUR	1
insurances	casualty insurances	3346	729541	EUR	1
credit institutions	public banks	8149	484449	EUR	2
insurances	life insurances	8715	247374	EUR	2
insurances	casualty insurances	17599	35040	EUR	2

In step 240 (see, e.g., FIG. 7C), the termination criterion for the loop i=i+1 for the remaining two data packages is checked anew (“not in parallel”). In this case, the comparison of data package 1 and data package 2 results in that the key of the first record of data package 2 is greater than the key of the first record of data package 1, and that the key of the last record of data package 1 is greater than said key of the first record of data package 2, which represents intersecting data packages. As a result, the termination criterion is not accomplished, consequently assigning the data package 1 to data package 2 for rebuilding a new data package 1. Since data package 1 includes 4 records and data package 2 includes 3 records, the data package size of the new data package 1 including M_m=7 records exceeds the maximum package size of M_p=4, which is acceptable.

In step 250 (see, e.g., FIG. 7C), the aggregated data packages are assigned again to the second parallel processing of the aggregation algorithm 30 of FIG. 2 illustrated by the step 50 within the scope of a third iteration (Iteration Nr. 3) for merging said adjacent data packages of step 230. Having merged the data package 1 and data package 2 to a new data package 1, the data records of the remaining new data package 1 are reorganized, and thereafter sorted again according to the given order for said granularity characteristics as illustrated in the table of FIG. 5 and in FIG. 4b, respectively. The outcome of this reorganization and sorting is illustrated hereinafter in Table 7.

TABLE 7
STEP: 250
		Sum	Minimum
		financial	financial
		statement key	statement key		old	new
rating method	rating-segment			currency	package	package
Kreditinstitute	private banks	1140	166310	EUR	1	1
Kreditinstitute	public banks	6114	456825	EUR	1	1
Kreditinstitute	public banks	8149	484449	EUR	2	1
Versicherungen	life insurances	7586	670365	EUR	1	1
Versicherungen	life insurances	8715	247374	EUR	2	1
Versicherungen	casualty insurances	3346	729541	EUR	1	1
Versicherungen	casualty insurances	17599	35040	EUR	2	1

Thereupon, the aggregation for said key figures using said predetermined aggregation operations is conducted just once more by assigning said data records to the first parallel process illustrated by method step 40 of FIG. 2 within the scope of a third iteration (Iteration Nr. 3), wherein as a result, the size of the new data package 1 decreases from M_m=7 to N_a=4 according to the customer defined granularity as illustrated in FIG. 5 referring to “search result”. Following the aggregation, the results of the remaining data packages are saved in a database. This outcome of this aggregation is illustrated hereinafter in Table 8.

TABLE 8
STEP: 250
		financial	financial
		statement key	statement key
rating method	rating-segment			currency	package
Kreditinstitute	private banks	1140	166310	EUR	1
Kreditinstitute	public banks	14263	456825	EUR	1
Versicherungen	casualty insurances	16301	247374	EUR	1
Versicherungen	casualty insurances	17599	35040	EUR	1

By checking the termination criterion for the loop i=i+1 once again in step 260 (“not in parallel”), the aggregation algorithm 30 of FIG. 2 terminates at this point, since there is no other adjacent data package, whose first key is an element of any other data package, or in other words interpreting the termination criterion, all the data packages are disjoint with regard to the granularity characteristics.

FIG. 6 illustrates an example of use for an optimized aggregation algorithm compared to the preceding example of use of FIG. 5, using a processing tool based on the parallel aggregation algorithm 30 of FIG. 2. The aggregation of the raw data illustrated in FIG. 4a including M=12 data records reduces the amount of data to 4<M=12 data records according to the customer defined granularity, as it is set out in FIG. 5 referring to “search result”.

The granularity fields including granularity characteristics are characterized by “rating method” and “rating segment”. The fields of key figures are characterized by the columns “financial statement key figure 1” and “financial statement key figure 2”.

The data package size is determined through customizing. According to the statement that large data package sizes are ideal for aggregating, whereas small data package seizes are ideal for reorganizing, in this example of use the data package size (M_p) for aggregating is determined relatively great with M_p=8 and the sub data package size (N_sp) is determined relatively low with N_sp=3, thereby complying with the interests working in opposite direction.

In step 200 (see, e.g., FIG. 8A), the raw data shown in the original work list of FIG. 4a are exemplarily arranged by the key figures in the column “financial statement key figure 1” in ascending order so as to demonstrate a work list of unsorted records to begin with. Due to the customized data package size of M_p=8, the M=12 data records of said work list are split in two data packages, in data package 1 for aggregating, including M_p=8 data records, and a remaining data package 2 corresponding to a remaining rest that is not to be aggregated, including 4 data records.

In analogy to the preceding example of use in FIG. 5, for the exemplification of the parallel aggregation algorithm as illustrated in FIG. 2 on the basis of the concrete example and to simplify matters, only the granularity fields characterized by “rating method” and “rating segment”, the fields of key figures characterized by the columns “financial statement key figure 1” and “financial statement key figure 2”, and the field currency are taken into consideration. All the rest of fields remain empty. Hereinafter, Table 9 illustrates the outcome of this reorganization and simplification of the original work list shown in FIG. 4a.

TABLE 9
STEP: 200
		Sum	Minimum
		financial	financial
		statement key	statement key
rating method	rating-segment			currency	package
insurances	life insurances	1620	865860	EUR	1
credit institutions	private banks	1912	809485	EUR	1
credit institutions	public banks	2860	456825	EUR	1
credit institutions	public banks	3254	693677	EUR	1
insurances	casualty insurances	3346	729541	EUR	2
credit institutions	private banks	3393	542616	EUR	2
insurances	life insurances	5966	670365	EUR	2
credit institutions	private banks	6135	166310	EUR	2
credit institutions	public banks	8149	484449	EUR	3
insurances	casualty insurances	8683	824001	EUR	3
insurances	life insurances	8715	247374	EUR	3
insurances	casualty insurances	8916	35040	EUR	3

In step 210 (see e.g., FIG. 8A), the data packages are assigned to the parallel processing of the aggregation algorithm 30 of FIG. 2. Within the scope of a first iteration (Iteration Nr. 1), the parallel processing begins with the method step 40 of FIG. 2. Up to this point, there is not identified one granularity level i mentioned above, which is symbolized by i=0 in FIG. 2. Therefore, at first, the fields of granularity characteristics labeled “rating method” and “rating segment” are identified so as to obtain levels i of granularity characteristics within said data packages, thereby accomplishing the criterion i>0. The maximum reachable number of granularity levels i per data package is i=4 due to the table illustrated in the preceding example of FIG. 5.

By sequentially comparing the customer defined granularity characteristics shown in the table of FIG. 5 with the data records of each of the two data packages, thereby beginning with the first row of granularity characteristics of the table of FIG. 5 characterized through “credit institutions/private banks” and ending with the forth row characterized through “insurances/casualty insurances”, the data records of data package 1 and data package 2 are searched for matching results. In our example of use there appear in total four matches in data package 1 and three matches in data package 2 in respect to said granularity characteristics and granularity levels i, respectively, what as a result corresponds to a granularity level of i=4 of identified granularity characteristics for data package 1 and i=3 for data package 2.

Subsequently, both of the data packages are sorted according to the given order as set out in the table of the preceding example of FIG. 5. The outcome of this sorting of data packages is illustrated below in Table 10.

TABLE 10
STEP: 200
		Sum	Minimum
		financial	financial
		statement key	statement key
rating method	rating-segment			currency	package
credit institutions	private banks	1912	809485	EUR	1
credit institutions	private banks	3393	542616	EUR	1
credit institutions	private banks	6135	166310	EUR	1
credit institutions	public banks	2860	456825	EUR	1
credit institutions	public banks	3254	693677	EUR	1
insurances	life insurances	1620	865860	EUR	1
insurances	life insurances	5966	670365	EUR	1
insurances	casualty insurances	3346	729541	EUR	1
credit institutions	public banks	8149	484449	EUR	2
insurances	life insurances	8715	247374	EUR	2
insurances	casualty insurances	8683	824001	EUR	2
insurances	casualty insurances	8916	35040	EUR	2

Subsequently, the rows of data package 1 are aggregated for the key figures (Xj) characterized through “financial statement key figure 1” and “financial statement key figure 2” by applying appropriate aggregation operations (operators j) to the respective key figures, wherein said aggregation operations being predetermined or customized aggregation operations. In this case the matches are added up in respect to key figure 1, and in respect to key figure 2 the minimum value is taken over, thereby reducing the number of data records. As a result, data package 1 is reduced to N_a=4<M_p=8 data records, which is illustrated below in Table 11.

TABLE 11
STEP: 210
		Sum	Minimum
		financial	financial
		statement key	statement key
rating method	rating-segment			currency	package
credit institutions	private banks	11440	166310 EUR	1
credit institutions	public banks	6114	456825	EUR 1
insurances	life insurances	7586	670365	EUR	1
insurances	casualty insurances3346	729541	EUR	1
credit institutions	public banks	8149	484449	EUR 2
insurances	life insurances	8715	247374	EUR	2
insurances casualty	insurances	8683	824001	EUR	2
insurances casualty	insurances	8916	35040	EUR	2

Thereafter, the data packages are split into sub data packages and then the sub data packages are saved in a database. Since the sub data package size (N_sp) is determined by N_sp=3, each of the two remaining data packages including 4 data records is split in two sub data packages, wherein each of the sub data packages 1 and 3 includes 3 data records, and each of the sub data packages 2 and 4 corresponding to the rest of data package 1 and data package 2, respectively, only includes 1 data record. The outcome of this splitting of data packages into sub data packages is illustrated below in Table 12.

TABLE 11
STEP: 210
		Sum	Minimum
		financial	financial
		statement key	statement key		old	new
rating method	rating-segment			currency	package	package
credit institutions	private banks	11440	166310	EUR	1	1
credit institutions	public banks	6114	456825	EUR	1	1
insurances	life insurances	7586	670365	EUR	1	1
insurances	casualty insurances	3346	729541	EUR	1	2
credit institutions	public banks	8149	484449	EUR	2	3
insurances	life insurances	8715	247374	EUR	2	3
insurances	casualty insurances	8683	824001	EUR	2	3
insurances	casualty insurances	8916	35040	EUR	2	4

Thereupon, in step 220 (see, e.g., FIG. 8B), the identification of adjacent data packages based on these small sub data packages is conducted by checking the termination criterion of the loop i=i+1 (“not in parallel”) being:

• • if key_pos1,xε(key_pos1,y;key_{pos max:y}) then continue else terminate,

wherein pos1 illustrates the first position of a data package, posmax illustrates the last position of a data package, and x, y illustrates the number of a data package, thereby comparing the key of the first record of each data package with the key of the first and the last record of all the rest of data packages (thus comparing all combinations x, y). If said criterion is not accomplished, meaning that the conditional inquiry is true, the data packages are assigned for rebuilding new data packages.

Beginning with the key of the first record of data package 1, the comparison of data package 1 and data package 2 results in that the key of the first record of data package 1 is less than the key of the single record of data package 2. As a result, interpreting the conditional inquiry for the loop i=i+1, data package 1 and data package 2 do not intersect. Accordingly, data package 1 and data package 4 do not intersect. In contrast, data package 1 and data package 3 appear to intersect, since the key of the first record of data package 2 is greater than the key of the first record of data package 1 and less than the key of the last record of data package 1, meaning that the key of the first record of data package 2 is an element of the amount of data records in data package 1. Thus, they are identified as adjacent data packages. Further, the keys of data package 2 and 4 are identical, and thus both packages are not disjoint. Accordingly, they are identified as adjacent data packages, too. Thereupon, the data packages identified as adjacent are assigned for rebuilding new data packages.

Since data package 1 and data package 2, respectively, only include 3 records, the data package size of the new data package 1 including 6 records is still less than the determined package size of M_p=8. The data package size of the new data package 2 includes 1+1=2 data records. In order to restore the original package size of M_p=8, the new data package 2 is additionally added to the new data package 1.

In step 230 (see, e.g., FIG. 8B), the sub data packages of step 210 are assigned to the second parallel processing of the aggregation algorithm 30 of FIG. 2 illustrated by the method step 50 within the scope of a second iteration (Iteration Nr. 2) for merging adjacent data packages and rebuilding new data packages, respectively. Thus, having merged sub data package 1 with sub data package 3, and sub data package 2 with sub data package 4, and additionally added the new data package 2 to the new data package 1, in all only one new data package remains. Subsequent to the merger, the data records are assigned again to the above mentioned first parallel process illustrated by step 40 of FIG. 2 within the scope of a second iteration (Iteration Nr. 2), wherein the data records of the remaining new data package 1 are reorganized, and thereafter sorted again according to the given order of said granularity characteristics as illustrated in the table of the preceding example of FIG. 5 and in FIG. 4b, respectively. The outcome of this reorganization and sorting is illustrated hereinafter in Table 13.

TABLE 13
STEP: 210
		Sum	Minimum
		financial	financial
		statement key	statement key		old	new
rating method	rating-segment			currency	package	package
credit institutions	private banks	11440	166310	EUR	1	1
credit institutions	public banks	6114	456825	EUR	1	1
credit institutions	public banks	8149	484449	EUR	3	1
insurances	life insurances	7586	670365	EUR	1	1
insurances	life insurances	8715	247374	EUR	3	1
insurances	casualty insurances	3346	729541	EUR	2	1
insurances	casualty insurances	8683	824001	EUR	3	1
insurances	casualty insurances	8916	35040	EUR	4	1

Thereupon, the aggregation for the key figures using said predetermined aggregation operations is conducted just once more, wherein as a result, the size of the new data package 1 decreases from M_p=8 to N_a=4 according to the customer defined granularity as illustrated in FIG. 5 referring to “search result”. This outcome of this aggregation is illustrated hereinafter in Table 14.

TABLE 14
STEP: 230
		financial	financial
		statement key	statement key
rating method	rating-segment			currency	package
credit institutions	private banks	11440	166310	EUR 1
credit institutions	public banks	14263	456825	EUR	1
insurances	life insurances	16301	247374	EUR	1
insurances	casualty insurances	20945	35040	EUR	1

Following the aggregation, the results of the remaining data package are saved in a database.

By checking the termination criterion for the loop i=i+1 once again in step 240 (see, e.g., FIG. 8C), the aggregation algorithm 30 of FIG. 2 terminates at this point, since there is no other adjacent data package, whose first key is an element of any other data package, or in other words interpreting the termination criterion, all the data packages are disjoint with regard to the granularity characteristics.

FIGS. 7A through 7C and FIGS. 8A through 8C show exemplary work lists or tables consistent with embodiments the present invention.

The foregoing description has been presented for purposes of illustration. It is not exhaustive and does not limit the invention to the precise forms or embodiments disclosed. Modifications and adaptations of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the disclosed embodiments of the invention. For example, the described implementations include software, but systems and methods consistent with the present invention may be implemented as a combination of hardware and software or in hardware alone. Examples of hardware include computing or processing systems, including personal computers, servers, laptops, mainframes, micro-processors and the like. Additionally, although aspects of the invention are described for being stored in memory, one skilled in the art will appreciate that these aspects can also be stored on other types of computer-readable media, such as secondary storage devices, for example, hard disks, floppy disks, or CD-ROM, the Internet or other propagation medium, or other forms of RAM or ROM.

Computer programs based on the written description and methods of this invention are within the skill of an experienced developer. The various programs or program modules can be created using any of the techniques known to one skilled in the art or can be designed in connection with existing software. For example, program sections or program modules can be designed in or by means of Java, C++, HTML, XML, or HTML with included Java applets or in SAP R/3 or ABAP. One or more of such software sections or modules can be integrated into a computer system or existing e-mail or browser software.

Moreover, while illustrative embodiments of the invention have been described herein, the scope of the invention includes any and all embodiments having equivalent elements, modifications, omissions, combinations (e.g., of aspects across various embodiments), adaptations and/or alterations as would be appreciated by those in the art based on the present disclosure. The limitations in the claims are to be interpreted broadly based on the language employed in the claims and not limited to examples described in the present specification or during the prosecution of the application, which examples are to be construed as non-exclusive. Further, the steps of the disclosed methods may be modified in any manner, including by reordering steps and/or inserting or deleting steps, without departing from the principles of the invention. It is intended, therefore, that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims and their full scope of equivalents.

Claims

1. A computer-implemented method for automated generic and parallel aggregation of characteristics and key figures of mass data, the mass data including M records from a single database of a single data source or from different databases of different data sources, particularly associated with financial institutions and with financial affairs in banking practice, and further including sets of rows and sets of columns, each row corresponding to a record, and the columns including fields of predetermined granularity characteristics and fields of predetermined key figures, wherein the aggregation reduces the amount of data to N≦M records for a customer defined granularity, the method comprising the following steps: receiving the mass data from a single database of a single data source or from different databases of different data sources associated with banking practice; selecting predetermined granularity characteristics and predetermined key figures, and selecting predetermined aggregation operations to be carried out by the processing means of a data processing system; reading input data from a single database of a single data source or from different databases of different data sources into the processing means of a data processing system; preparing the input data as data packages being of the size Mp in a preparational step before the aggregation starts; processing the data packages being of the size Mp in a parallel process by identifying the granularity characteristics, thereby identifying unique granularity levels i; sorting the records of each data package for a given order of granularity characteristics of the customized granularity; and subsequently aggregating the records in each data package for key figures by using aggregation operations; and following the aggregation, saving the results of each data package.

2. The method of claim 1, wherein the aggregation is computed for the pre-determined granularity characteristics and/or predetermined key figures using predetermined aggregation operations selected from a function pool and/or costumer defined aggregation operations to be defined by input means using the predetermined aggregation operations.

3. The method of claim 1, wherein the aggregation is computed for costumer defined granularity characteristics and/or costumer defined key figures that are to be defined by input means using the predetermined aggregation operations selected from a function pool and using the predetermined aggregation operations and/or the costumer defined aggregation operations.

4. The method of claim 1, wherein the data packages being of the size Mp are processed within a loop i=i+1 comprising the steps of: a first parallel process for identifying the granularity characteristics, thereby identifying unique granularity levels i; sorting the records of each data package for a given order of granularity characteristics of the customized granularity; and subsequently aggregating the data records in each data package for key figures by using aggregation operations; thereby reducing the amount of data records to Na≦Mp; and following the aggregation, saving the results of each data package in a local database and storing the size and the key of the first and the last record of each data package in a global database; and subsequently checking the termination criterion for the loop i=i+1 (“not in parallel”) being: if keypos1,xε(keypos1,y; keypos max:y) then continue else terminate, wherein pos1 illustrates the first position of a data package, posmax illustrates the last position of a data package, and x, y illustrates the number of data package, and if the conditional criterion is not accomplished for all combinations (x, y), meaning that the conditional inquiry is true, thereby comparing the key of the first record of each data package with the first and the last record of all the rest of packages, assigning the aggregated packages to a second parallel process for merging adjacent data packages so as to rebuild new data packages, wherein adjacent packages are those packages with keys of the first record which are closest together and have violated the termination criterion, then storing the merged packages to a local database, and subsequently assigning the merged data packages again to the above mentioned first parallel process for reorganizing and sorting the new data packages, and thereafter aggregating the new data packages for key figures by using aggregation operations, and following the aggregation, after each loop cycle checking the termination criterion for the loop i=i+1 for all combinations (x, y) anew, repeatedly executing the loop i=i+1 while the termination criterion for the loop is not accomplished, else after accomplishing the criterion, i.e. all the data packages are disjoint with regard to the granularity characteristics, terminating the loop i=i+1.

5. The method of claim 1, wherein the data packages being of the size Mp are processed within a loop i=i+1 comprising the steps of: a first parallel process for identifying the granularity characteristics, thereby identifying unique granularity levels i; sorting the records of each data package for a given order of granularity characteristics of the customized granularity; and subsequently aggregating the data records in each data package for key figures by using aggregation operations; thereby reducing the amount of data records to Na≦Mp, and following the aggregation, splitting the aggregated data packages into several smaller data sub packages being of the size Nsp and saving the results of each sub data package in a local database; storing the size and the key of the first and the last record of each sub data package to a global database; and subsequently identifying adjacent packages based on these small sub data packages by checking (“not in parallel”) the termination criterion for the loop i=i+1 being: if keypos1,xε(keypos1,y; keypos max:y) then continue else terminate, wherein pos1 illustrates the first position of a data package, posmax illustrates the last position of a data package, and x, y illustrates the number of data package, and if the conditional criterion is not accomplished for all combinations (x, y), meaning that the conditional inquiry is true, thereby comparing the key of the first record of each sub data package with the first and the last record of all the rest of sub data packages, assigning the sub data packages to a second parallel process for merging adjacent sub data packages so as to rebuild new data packages, wherein adjacent sub data packages are those data packages with keys of the first record which are closest together and have violated the termination criterion, and wherein by merging the sub data packages the original package size N is restored; then storing the new data packages to a local database; and subsequently assigning the new data packages again to the above mentioned first parallel process for reorganizing and sorting; and thereafter aggregating the new data packages for key figures by using aggregation operations; and following the aggregation, after each loop cycle splitting the aggregated data packages again into several smaller sub data packages and saving the results of each sub data package in a local database; storing the size and the key of the first and the last record of each sub data package to a global database; and subsequently identifying adjacent packages again based on these small packages by checking (“not in parallel”) the termination criterion for the loop i=i+1 for all combinations (x, y) anew, repeatedly executing the loop i=i+1 while the termination criterion for the loop is not accomplished, else after accomplishing the criterion, i.e. all the data packages are disjoint with regard to the granularity characteristics, terminating the loop i=i+1.

6. The method of claim 4, wherein ultimately conducting an additional calculation step for enriching aggregated data packages, and subsequently saving the data packages to a global result database.

7. The method of claim 1, further comprising the steps of: enriching the prepared data packages in a parallel pre-processing step via secondary data source or data sources before the parallel aggregation starts; and saving the results to a local database.

8. The method of claim 6, further comprising the steps of: enriching the aggregated data packages in a parallel post-processing step via secondary data source or data sources following the parallel aggregation; and saving the results to a global result database.

9. The method of claim 1, wherein the data packages are processed within jobs, each of the jobs including a plurality of data packages.

10. The method of claim 9, wherein one job or a plurality of jobs are processed in a parallel processing mode using a single data processor.

11. The method of claim 9, wherein one job or a plurality of jobs are processed in a parallel processing mode using a network of data processors.

12. The method of claim 11, wherein the network of data processors is a Local Area Network (LAN), Wide Area Network (WAN), intranet or internet.

13. The method of claims 1, wherein the data packages are processed within jobs, and wherein the jobs are processed in a parallel processing mode using a single data processor, thereby aggregating and merging the data packages of a job sequentially.

14. The method of claim 1, wherein the data packages are processed within jobs, and wherein the jobs are processed in a parallel processing mode using a network of data processors, thereby aggregating and merging the data packages of a job sequentially.

15. A computer system configured to perform automated generic and parallel aggregation of characteristics and key figures of mass data, the mass data including M records from a single database of a single data source or from different databases of different data sources, particularly associated with financial institutions and with financial affairs in banking practice, and further including sets of rows and sets of columns, each row corresponding to a record, and the columns including fields of predetermined granularity characteristics and fields of predetermined key figures, wherein the aggregation reduces the amount of data to N≦M records for a customer defined granularity, comprising: means for receiving the mass data from a single database of a single data source or from different databases of different data sources associated with banking practice; means for selecting predetermined granularity characteristics and predetermined key figures, and selecting predetermined aggregation operations to be carried out by the processing means of a data processing system; means for reading input data from a single database of a single data source or from different databases of different data sources into the processing means of a data processing system; means for preparing the input data as data packages being of the size Mp in a preparational step before the aggregation starts; means for processing the data packages being of the size Mp in a parallel process by identifying the granularity characteristics, thereby identifying unique granularity levels i; sorting the records of each data package for a given order of granularity characteristics of the customized granularity; and subsequently aggregating the records in each data package for key figures by using aggregation operations; and means for, following the aggregation, saving the results of each data package.

16. A computer system of claim 15, wherein the processing means is configured to process the data packages being of the size Mp within a loop i=i+1 and includes the steps of: a first parallel process for identifying the granularity characteristics, thereby identifying unique granularity levels i; sorting the records of each data package for a given order of granularity characteristics of the customized granularity; and subsequently aggregating the data records in each data package for key figures by using aggregation operations; thereby reducing the amount of data records to Na≦Mp; and following the aggregation, saving the results of each data package in a local database and storing the size and the key of the first and the last record of each data package in a global database; and subsequently checking the termination criterion for the loop i=i+1 (“not in parallel”) being: if keypos1,xε(keypos1,y; keypos max:y) then continue else terminate, wherein pos1 illustrates the first position of a data package, posmax illustrates the last position of a data package, and x, y illustrates the number of data package, and if the conditional criterion is not accomplished for all combinations (x, y), meaning that the conditional inquiry is true, thereby comparing the key of the first record of each data package with the first and the last record of all the rest of packages, assigning the aggregated packages to a second parallel process for merging adjacent data packages so as to rebuild new data packages, wherein adjacent packages are those packages with keys of the first record which are closest together and have violated the termination criterion, then storing the merged packages to a local database, and subsequently assigning the merged data packages again to the above mentioned first parallel process for reorganizing and sorting the new data packages, and thereafter aggregating the new data packages for key figures by using aggregation operations, and following the aggregation, after each loop cycle checking the termination criterion for the loop i=i+1 for all combinations (x, y) anew, repeatedly executing the loop i=i+1 while the termination criterion for the loop is not accomplished, else after accomplishing the criterion, i.e. all the data packages are disjoint with regard to the granularity characteristics, terminating the loop i=i+1.

17. A computer system of claim 15, wherein the processing means are configured to process the data packages being of the size Mp within a loop i=i+1 and includes the steps of: a first parallel process for identifying the granularity characteristics, thereby identifying unique granularity levels i; sorting the records of each data package for a given order of granularity characteristics of the customized granularity; and subsequently aggregating the data records in each data package for key figures by using aggregation operations; thereby reducing the amount of data records to Na≦Mp and following the aggregation, splitting the aggregated packages into several smaller sub packages and saving the results of each sub data package in a local database; storing the size and the key of the first and the last record of each sub data package to a global database; and subsequently identifying adjacent packages based on these small sub data packages by checking (“not in parallel”) the termination criterion for the loop i=i+1 being: if keypos1,xε(keypos1,y; keypos max:y) then continue else terminate, wherein pos1 illustrates the first position of a data package, posmax illustrates the last position of a data package, and x, y illustrates the number of data package, and if the conditional criterion is not accomplished for all combinations (x, y), meaning that the conditional inquiry is true, thereby comparing the key of the first record of each sub data package with the first and the last record of all the rest of sub data packages, assigning the sub data packages to a second parallel process for merging adjacent sub data packages so as to rebuild new data packages, wherein adjacent sub data packages are those data packages with keys of the first record which are closest together and have violated the termination criterion, and wherein by merging the sub data packages the original package size N is restored; then storing the new data packages to a local database; and subsequently assigning the new data packages again to the above mentioned first parallel process for reorganizing and sorting; and thereafter aggregating the new data packages for key figures by using aggregation operations; and following the aggregation, after each loop cycle splitting the aggregated data packages again into several smaller sub data packages and saving the results of each sub data package in a local database; storing the size and the key of the first and the last record of each sub data package to a global database; and subsequently identifying adjacent packages again based on these small packages by checking (“not in parallel”) the termination criterion for the loop i=i+1 for all combinations (x, y) anew, repeatedly executing the loop i=i+1 while the termination criterion for the loop is not accomplished, else after accomplishing the criterion, i.e. all the data packages are disjoint with regard to the granularity characteristics, terminating the loop i=i+1.

18. A computer system of claim 16, wherein the processing means is further configured to ultimately conduct an calculation step for enriching data packages, and wherein the storage means are further configured to subsequently save the data packages to a global result database.

19. A computer system of claim 15, wherein the processing means is further configured to ultimately enrich the prepared data packages in a parallel pre-processing step via secondary data source or data sources before the parallel aggregation starts, and wherein the storage means are further configured to save the results to the local database.

20. A computer system of claim 18, wherein the processing means is further configured to ultimately enrich the aggregated data packages in a parallel post-processing step via secondary data source or data sources following the parallel aggregation, and wherein the storage means are further configured to save the results to the global database.

21. A computer program product having a plurality of instructions for causing processing means of a computer system to execute steps of: receiving the mass data from a single database of a single data source or from different databases of different data sources associated with banking practice; selecting predetermined granularity characteristics and predetermined key figures, and selecting predetermined aggregation operations to be carried out by the processing means of a data processing system; reading input data from a single database of a single data source or from different databases of different data sources into the processing means of a data processing system; preparing the input data as data packages being of the size Mp in a preparational step before the aggregation starts; processing the data packages being of the size Mp in a parallel process by identifying the granularity characteristics, thereby identifying unique granularity levels i; sorting the records of each data package for a given order of granularity characteristics of the customized granularity; and subsequently aggregating the records in each data package for key figures by using aggregation operations; and following the aggregation, saving the results of each data package.

22. The computer program product of claim 21, wherein the program comprises instructions for processing the data packages being of the size Mp within a loop i=i+1 including the steps of: a first parallel process for identifying the granularity characteristics, thereby identifying unique granularity levels i; sorting the records of each data package for a given order of granularity characteristics of the customized granularity; and subsequently aggregating the data records in each data package for key figures by using aggregation operations; thereby reducing the amount of data records to Na≦Mp; and following the aggregation, saving the results of each data package in a local database and storing the size and the key of the first and the last record of each data package in a global database; and subsequently checking the termination criterion for the loop i=i+1 (“not in parallel”) being: if keypos1,xε(keypos1,y; keypos max:y) then continue else terminate, wherein pos1 illustrates the first position of a data package, posmax illustrates the last position of a data package, and x, y illustrates the number of data package, and if the conditional criterion is not accomplished for all combinations (x, y), meaning that the conditional inquiry is true, thereby comparing the key of the first record of each data package with the first and the last record of all the rest of packages, assigning the aggregated packages to a second parallel process for merging adjacent data packages so as to rebuild new data packages, wherein adjacent packages are those packages with keys of the first record which are closest together and have violated the termination criterion, then storing the merged packages to a local database, and subsequently assigning the merged data packages again to the above mentioned first parallel process for reorganizing and sorting the new data packages, and thereafter aggregating the new data packages for key figures by using aggregation operations, and following the aggregation, after each loop cycle checking the termination criterion for the loop i=i+1 for all combinations (x, y) anew, repeatedly executing the loop i=i+1 while the termination criterion for the loop is not accomplished, else after accomplishing the criterion, i.e. all the data packages are disjoint with regard to the granularity characteristics, terminating the loop i=i+1.

23. The computer program product of claim 21, wherein the program comprises instructions for processing the data packages being of the size Mp within a loop i=i+1 including the steps of: a first parallel process for identifying the granularity characteristics, thereby identifying unique granularity levels i; sorting the records of each data package for a given order of granularity characteristics of the customized granularity; and subsequently aggregating the data records in each data package for key figures by using aggregation operations; thereby reducing the amount of data records to Na≦Mp, and following the aggregation, splitting the aggregated data packages into several smaller data sub packages being of the size Nsp and saving the results of each sub data package in a local database; storing the size and the key of the first and the last record of each sub data package to a global database; and subsequently identifying adjacent packages based on these small sub data packages by checking (“not in parallel”) the termination criterion for the loop i=i+1 being: if keypos1,xε(keypos1,y; keypos max:y) then continue else terminate, wherein pos1 illustrates the first position of a data package, posmax illustrates the last position of a data package, and x, y illustrates the number of data package, and if the conditional criterion is not accomplished for all combinations (x, y), meaning that the conditional inquiry is true, thereby comparing the key of the first record of each sub data package with the first and the last record of all the rest of sub data packages, assigning the sub data packages to a second parallel process for merging adjacent sub data packages so as to rebuild new data packages, wherein adjacent sub data packages are those data packages with keys of the first record which are closest together and have violated the termination criterion, and wherein by merging the sub data packages the original package size N is restored; then storing the new data packages to a local database; and subsequently assigning the new data packages again to the above mentioned first parallel process for reorganizing and sorting; and thereafter aggregating the new data packages for key figures by using aggregation operations; and following the aggregation, after each loop cycle splitting the aggregated data packages again into several smaller sub data packages and saving the results of each sub data package in a local database; storing the size and the key of the first and the last record of each sub data package to a global database; and subsequently identifying adjacent packages again based on these small packages by checking (“not in parallel”) the termination criterion for the loop i=i+1 for all combinations (x, y) anew, repeatedly executing the loop i=i+1 while the termination criterion for the loop is not accomplished, else after accomplishing the criterion, i.e. all the data packages are disjoint with regard to the granularity characteristics, terminating the loop i=i+1.

24. The computer program product of claim 22, wherein the program ultimately conducts an additional calculation step for enriching aggregated data packages, and wherein the data packages are subsequently saved to a global result database.

25. The computer program product of claim 21, further comprising the steps of: enriching the prepared data packages in a parallel pre-processing step via secondary data source or data sources before the parallel aggregation starts; and saving the results to a local database.

26. The computer program product of claim 24, further comprising the steps of: enriching the aggregated data packages in a parallel post-processing step via secondary data source or data sources following the parallel aggregation; and saving the results to a global result database.