JP2000067135A - Multi-dimensional analytic system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To automatically prepare a literal table based on the table of an accumulation source and to prepare accumulation data at high speed. SOLUTION: A literal is an index written on a slip, for example. The literal table shows a hierarchical relation between literal and is stored in a literal table storage part. A literal table generating mechanism generates the literal table based on the table of the accumulation source. The accumulation data show the accumulation of amounts or prices, for example. An accumulation data generating mechanism writes the accumulation data in the cells of an accumulation matrix based on the table of the accumulation source and the literal table. On the assumption that a shop Y corresponds to a row (j) and the year of '1997' corresponds to a column (k), for example, the sale at the shop Y in the year of 1997 is written in the cells on the row (j) and column (k) of the accumulation matrix. The accumulation matrix is stored in an accumulation data storage part.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multidimensional analysis system and, more particularly, to the creation of a literal table and aggregate data constituting a multidimensional database of the multidimensional analysis system.

[0002]

2. Description of the Related Art FIG. 10 is a diagram showing a configuration example of a multidimensional analysis system. In the figure, 1 is a display / operation unit, 2 is a control unit, 3 is a display data generation unit, 4 is a display information access unit, 5 is a display information storage unit, 6 is a DB access unit, 7 is a multidimensional database, 8 Denotes a total data storage unit, and 9 denotes a literal information storage unit. DB is an abbreviation for database. In addition, about the multidimensional analysis system,
For example, 1998 Patent Application No. 185394 or PFU
・ Technical Review 1998-May issue VOL.
9, NO. 1, Vol. 15, No. 10, pp. 10-17.

[0005] The display / operation unit 1 exists on, for example, a personal computer. The control unit 2, the display data generation unit 3, the display information access unit 4, the display information storage unit 5, the DB access unit 6, and the database 7 exist on, for example, a server. The personal computer and the server are connected by a LAN or the like. The total data storage unit 8 and the literal information storage unit 9 exist in the database 7. The control unit 2, the display data generation unit 3, the display information access unit 4, and the DB access unit 6 are realized by a program.

The display / operation unit 1 performs an analysis operation for displaying display data of the form created by the display data generation unit 3 and displaying detailed display information. The control unit 2 controls access to form display information, access to the DB and generation of form display data, and also controls result display and operation of the form.

[0005] The display data generator 3 displays the display information and D
Form display data is created based on the information of B. The generation of the detailed display information at the time of the analysis operation is also performed by the display data generation unit 3.
Done in The display information access unit 4 accesses a display information storage unit 5 that stores display information indicating a form format. The DB access unit 6 accesses the total data storage unit 8 and the literal information storage unit 9.

The display information stored in the display information storage unit 5 has information constituting a format of a form. FIG.
Describes the display information. The information that constitutes the display information includes axis information, hierarchy information, item information, data information, overall format information, and the like. An example of the display information is shown in FIG.

The axis information indicates the horizontal axis and the vertical axis, and an ID is assigned to each axis. For example, the axis ID of the vertical axis is 1, and the axis ID of the horizontal axis is 2. The hierarchy information indicates the hierarchy on the horizontal axis and the vertical axis, and an ID is assigned to each hierarchy. For example, the first hierarchical level I
D is 1 and the layer ID of the second layer is 2.

[0008] The item information indicates an item when displaying the total data, and indicates a literal for each item. FIG.
For example, a character string “quantity” is written in the location of item 1 of item 1, and a character string of “amount” is written in the location of item 2 in FIG. The data information represents an area for displaying the item value, and has a display format of the item value to be displayed in the data for each item. The display format indicates, for example, with / without digit separation and with / without currency symbol.

[0009] Information on the entire format includes a display position and a display attribute. FIG. 12 is a diagram illustrating a display position.
The display position indicates a display position of a literal displayed on the vertical axis and the horizontal axis (including items). The display attribute indicates an attribute of each item to which a display position ID is assigned. The display attributes include, for example, displaying a DB literal, displaying an item literal, and the like.

The multidimensional database 7 has a total data storage unit 8 and a literal information storage unit 9. The tally data stored in the tally data storage unit 8 has information constituting the tally data of the form. FIG. 13 is a diagram illustrating the aggregate data. The information that constitutes the aggregated data includes vertical codes,
There are horizontal codes and items. The aggregate data is composed of a plurality of rows. Note that the top row is a header row.
Each line excluding the header line has a vertical code field, a horizontal code field, an item 1 field, an item 2 field,..., An item m field, and the like. The value of the vertical code is entered in the field of the vertical code, and the value of the horizontal code is entered in the field of the horizontal code.
Item values are entered in the fields (i is 1,..., M).

The vertical code is link information with the literal information. When displaying the value of the item included in the line having the vertical code, the vertical code literal to be displayed simultaneously can be found from the literal information storage unit. Used for The horizontal code is the link information with the literal information, and is used to find the horizontal axis literal displayed at the same time from the literal information storage unit when displaying the value of the item included in the line having the horizontal code. You.

FIG. 14 is a view for explaining literal information. The literal information has information constituting a display literal of the form for each axis. Literal information is composed of a plurality of lines. Note that the top row is a header row. Each line, excluding the header line, contains a code field, a child number field,
It has a literal field and so on. In the code field, a code such as "A01" is written, and in the number of children field, the number of literals displayed in the lower hierarchy of the literal corresponding to the code is written.
A literal such as "1997" is written in the literal field.

FIG. 15 is a view for explaining codes. "A
The code “01” corresponds to the first literal of the first layer of the axis, and the code “A02” corresponds to the first literal of the axis.
It corresponds to the second literal in the hierarchy, and "B01"
The code “B02” corresponds to the first literal of the second layer of the axis, and the code “B02” corresponds to the second literal of the second layer of the axis, and is “B03”. The code corresponds to the third literal of the second layer of the axis, and the code “B04” corresponds to the fourth literal of the second layer of the axis. In the illustrated example, a literal of item 1 and a literal of item 2 are arranged below each of the lowermost literals on the horizontal axis.

FIGS. 16 and 17 are diagrams for explaining examples of form data generation. Note that (1) and (2) in FIGS.
,... Correspond to the following (1), (2),. (1) Obtain display information for each cell according to the hierarchy to be displayed. Note that a cell means a display unit, and each row of display information (excluding a header row) corresponds to one cell. (2) Sort the cells according to the order of the display positions of the display information. (3) Extract the hierarchical information to be displayed from the literal information,
Place code and literals. (4) After performing the above processing on the vertical and horizontal axes, a matrix is created by combining these. (5) From the aggregated data, extract the item values that match the display hierarchy and apply them to the matrix.

[0015]

FIG. 18A shows a table of a tallying source, and FIG. 18B shows an example of a form created from the table of a tallying source. In the prior art, when it is necessary to add, for example, "TV36" of "TV" in the table of the tallying source in FIG. 2A, the literal information has to be changed. In addition, when creating total data composed of multiple dimensions and multiple layers, it is necessary to total (lead) the table of the total source for each layer.

The present invention has been made in view of this point, and it is possible to automatically create a literal table based on a table of a tallying source, and to perform totaling based on a table of a tallying source and a literal table. The purpose is to be able to create data at high speed.

[0017]

FIG. 1 is a diagram illustrating the principle of the present invention. 2. The multidimensional analysis system according to claim 1, wherein a literal written on the form, a hierarchy indicating the hierarchy in which the literal is written, a code indicating the number of the literal in the hierarchy, a parent code, and a literal indicating the number of children. A multidimensional analysis system comprising: a literal table storage unit that stores a literal table having information for each of a plurality of axes; a total data storage unit that stores total data written on a form; and a literal table generation mechanism The literal table generation mechanism reads out one record from the table of the aggregation source, and checks whether a literal included in the read out record exists on the literal table of the corresponding axis. If not present, the literal information corresponding to the literal is stored in the literal table of the relevant axis. While pressurized, it is characterized in that performing the processing for correcting the literal information of a parent literal of the literal.

A multidimensional analysis system according to a second aspect of the invention provides a literal written on a form, a code indicating a hierarchy in which the literal is written, a code indicating the number of the literal in the hierarchy, a parent code, and the number of children. Table storage unit that stores a literal table having literal information indicating each of a plurality of axes, a total data storage unit that stores a total matrix in which total data written on a form is written, and a total data generation mechanism A multi-dimensional analysis system comprising: a total data generating mechanism, wherein a column corresponds to a literal on the vertical axis, and a row generates a total matrix corresponding to the literal on the horizontal axis; and The process of reading one record and the horizontal axis literal specified by the read record while referring to the literal table And a process of searching for a cell of an aggregation matrix corresponding to a literal on the vertical axis and a process of increasing the numerical value of the searched cell by the numerical value included in the read record. .

[0019]

FIG. 2 is a diagram showing an example of a literal tree structure. The literal tree structure logically shows the relationship between literals. A literal tree structure exists for each of the vertical axis, horizontal axis, and slice axis.

In the example shown in the figure, the node belonging to the highest layer has its own code of A1, and its parent code is 0.
(Indicating that there is no parent) and information indicating that the subordination (the number of children) is 2. The first node belonging to the middle layer has information indicating that its own code is B1, the parent code is A1, and the number of children is 6, and the like. The first node belonging to the lowest layer has information indicating that its own code is C1, the parent code is B1, and the number of children is 0, and the like. The literal tree structure is physically stored in storage in the form of a literal table.

FIG. 3 is a diagram for explaining a method of generating a literal table. Processing for generating a literal table will be described. Read the data of the table of the aggregation source. Check matching with each layer. If they match, the literal attribute is changed from "0" to "1". In the case of unmatch, after creating a chain structure (chain at the end and set the literal attribute to “2”), the corresponding hierarchical information is registered. At this time, if the newly created literal table is located at a lower level, the code of the corresponding upper-level literal table is set to its own “parent code”. Also, the subordinate number of the corresponding upper-level literal table is increased by one. The above processes are repeated until EOF. After the above processing is completed, the same processing is performed for the vertical axis and the slice axis.

The generation of a literal table will be described in detail with reference to FIG. The top row of the literal table (supposed to be row 0) is the header row. The rows excluding the header row have an axis ID field, a code field, a literal field, a parent code field, a dependent number field, a counting position field, and a literal attribute field.

In the first row of the vertical axis uppermost literal table, "Vertical, A00001, Taniguchi Electric, 0, 1, 0, 2"
There is literal information such as
In the first row of the table, "Vertical, B00001, TV 33
It is assumed that literal information “type, A00001,0,0,2” exists. In this state, when one record is taken out from the table of the counting source, the literal of one layer on the vertical axis of the record is “Other”, and
Suppose the literal of the hierarchy was "PC".

Since the literal on the first layer on the vertical axis of the retrieved record is "others", it is checked whether or not there is a row whose literal field is "others" in the uppermost literal table on the vertical axis. . Since there is no such case in this case, the line “Vertical, A00002, Other, 0, 0, 0, 2” is added to the second line of the vertical axis uppermost literal table.

Since the literal on the second layer on the vertical axis of the retrieved record is “PC” having “other” as a parent,
It is checked whether or not there is a line in the literal table of the second layer on the vertical axis, in which the literal field is “PC” and the parent code field is “A00002” corresponding to “Other”. In this case it doesn't exist,
"Vertical, B00002, PC, A00002, 0,
The line “0, 2” is added to the second line of the literal table in the second layer of the vertical axis. Further, the value of the “dependent number” existing in the second row of the vertical axis uppermost literal table is changed from 0 to 1.

FIGS. 4 and 5 are diagrams for explaining the high-speed aggregation of data. Aggregation creates an aggregation matrix area,
It is realized by adding. The algorithm is shown below. Data mart table expansion area (aggregation matrix area)
To win. The illustrated example is a two-dimensional example. If it becomes three-dimensional, a plurality of two-dimensional aggregation matrix areas are created. Data is read from the table of the aggregation source, and the literal
While referring to the table, add the value to the cell of the applicable aggregation matrix area.

The high-speed aggregation of data will be described in detail with reference to FIGS. First, an aggregation matrix area is obtained. The vertical size and the horizontal size of the aggregation matrix area are predetermined.

The source table is a relational database. In the illustrated example, the value of the first (leftmost) field of the record is represented on the horizontal axis, the value of the third field is represented on the vertical axis as a literal in the uppermost layer, and the value of the fifth field is represented on the vertical axis 2 The value of the sixth field is the literal of the top layer of the slice axis, and the value of the seventh field is the literal of the second layer of the slice axis.

Data is read from the table of the tallying source, and values are added to the cells of the applicable tallying matrix area while referring to the literal table. In the example shown in the figure, the record extracted from the counting source table is “9701,1, Taniguchi Electric, TV1, TV33, Overall, Tokyo, 1,1”.
00 ”.

Assume that the vertical axis top literal table is as shown in FIG. Also, the second vertical axis literal table is the same as the second vertical axis literal table shown in FIG. 3 except that a line “Vertical, B00003, TV 29 type, A00001,...” Is added at the end. Assume.

Then, the first row (numbered “1”) of the aggregation matrix area corresponds to “Taniguchi Denki”, the second row of the aggregation matrix area corresponds to “Other”, and The third row of the matrix area corresponds to “TV 33”, the fourth row of the tally matrix area corresponds to “TV 29”, and the fifth row of the tally matrix area corresponds to “PC”.

The literal belonging to the uppermost layer on the horizontal axis is “19”.
1997 "and" 1998 "," 1997 ",
Under each of "1998" there are child literals called "1st half" and "2nd half", and under "1st half", "January",
There is a child literal “February”,…, “June”
It is assumed that child literals “July”, “August”,..., “December” exist under “second half”.

Then, the first column (numbered “1”) of the aggregation matrix area is “1997”.
Year, and the second column of the aggregation matrix area is “199
Eight years ", and the third column of the aggregation matrix area is" 19
The fourth column of the aggregation matrix area corresponds to “2nd half of 1997”, the fifth column of the aggregation matrix area corresponds to “1st half of 1998”, and the sixth column of the aggregation matrix area. The column corresponds to “2nd half of 1998”, the seventh column of the aggregation matrix area corresponds to “January 1997”, and the eighth column of the aggregation matrix area corresponds to “2nd year of 1997”.
Corresponding to “month”, and the 29th column of the aggregation matrix area is “19”.
"November 1998", which corresponds to 30
The column corresponds to "December 1998".

The first column of the aggregation matrix area corresponds to 1997, the third column corresponds to the first half of 1997, the seventh column corresponds to January 1997, and the first row is “Taniguchi Denki”.
, And the third line corresponds to “TV 33”. Thus, the quantity of the retrieved record is represented by the cell in the first row and the first column, the cell in the first row and the third column, the cell in the first row and the seventh column, the cell in the third row and the first column, the cell in the third row and the third column, It is added to the cell in the third row and the seventh column.

FIG. 6 is a diagram showing a configuration example of a literal table automatic generation mechanism. In the figure, reference numeral 10 denotes a table of a totaling source, 11 denotes a DB access unit, 12 denotes a literal table automatic generation mechanism, 13 denotes a data reading unit, 14 denotes a matching processing unit, 15 denotes a tree structure creating unit, and 16 denotes a literal table. A sort unit 17 indicates a literal table control unit.

The automatic literal table generation mechanism 12
It has a data reading unit 13, a matching processing unit 14, a tree structure creating unit 15, a literal table sorting unit 16, and a literal table control unit 17. The data reading unit 13 issues a record read request for extracting one record from the table of the counting source. When a record reading request is issued, the DB access unit 11 retrieves one record from the table 10 of the tallying source, and passes the retrieved record to the data reading unit 13. The data reading unit 13 passes the received record to the matching processing unit 14.

The matching processing section 14 performs the matching processing as described in the method of generating a literal table, and activates the tree structure creating section 15 in the case of an unmatch. The tree structure creation unit 15 creates a row including new literal information, adds the created row to the corresponding literal table, and activates the literal table sorting unit 16. The literal table sorting unit 16 performs a process of writing a parent code in a row newly added to the literal table or increasing the number of subordinations in a row that becomes a parent of the added row by one. The literal table control unit 17 performs a process for controlling the entire literal table automatic generation mechanism.

FIG. 7 is a diagram showing a processing flow for generating a literal table. In step S1, one record is read from the table of the tallying source. If it is not EOF (end of file), the process proceeds to step S2, and if it is EOF, the process proceeds to step S2. In step S2, it is checked whether the same literal as the literal included in the read record exists on the literal table. If it exists, the process returns to step S1. In step S3, tree information is added to the literal table. Further, a process of creating a parent-child relationship and increasing the number of parents is performed.

FIG. 8 is a diagram showing a configuration example of the data high-speed aggregation mechanism. In the figure, 18 is a data high-speed aggregation mechanism,
19 is a data reading unit, 20 is a tally matrix creation unit,
21 is a corresponding cell search unit, 22 is a data addition unit, 23
Indicates a literal table control unit. In addition, the same code | symbol as FIG. 6 shows the same thing.

The data high-speed aggregation mechanism 18 has a data reading unit 19, an aggregation matrix creation unit 20, a corresponding cell search unit 21, a data addition unit 22, and a literal table control unit 23. The data reading unit 19 issues a record reading request. When a record reading request is issued, the DB access unit 11 reads a record from the table of the tallying source, and passes the read record to the data reading unit 19.

The aggregation matrix creating section 20 creates an aggregation matrix. The corresponding cell search unit 20 refers to the literal table and searches for a cell in the aggregation matrix corresponding to the literal included in the record passed from the data reading unit 19. The data adding unit 22 adds a value present in the record passed from the data reading unit 19 to the value of the cell searched by the corresponding cell searching unit 21 and writes the addition result back to the original cell. The literal table control unit 23 performs a process for controlling the entire data high-speed aggregation mechanism.

FIG. 9 is a diagram showing a processing flow of data high-speed aggregation. In step S1, an aggregation matrix is created. In step S2, one record is read from the table of the counting source. In the case of EOF, the process is terminated, and in the case of not EOF, the process proceeds to step S3. In step S3, a cell on the aggregation matrix corresponding to the read record is searched for with reference to the literal table. Step S
In step 4, data is added to the corresponding cell. Data addition is
This is performed for the corresponding cells. Next, the process returns to step S2.

[0043]

As is apparent from the above description, according to the present invention, a literal table can be automatically generated based on a table of a totaling source. Therefore, when, for example, the data of “TV 36” is added to the table of the counting source, the literal information of “TV 36” can be added to the literal table. According to the present invention,
Aggregated data can be created with a single lead for the table of the aggregation source, and the creation time of the aggregated data can be reduced.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is a diagram illustrating an example of a literal tree structure.

FIG. 3 is a diagram for explaining a method of generating a literal table.

FIG. 4 is a diagram for explaining high-speed aggregation of data.

FIG. 5 is a diagram for explaining high-speed aggregation of data (continued)
It is.

FIG. 6 is a diagram illustrating an example of a literal table automatic generation mechanism.

FIG. 7 is a diagram showing a processing flow of literal table generation.

FIG. 8 is a diagram illustrating a configuration example of a data high-speed aggregation mechanism.

FIG. 9 is a diagram illustrating a processing flow of data high-speed aggregation.

FIG. 10 is a diagram illustrating a configuration example of a multidimensional analysis system.

FIG. 11 is a diagram illustrating display information.

FIG. 12 is a diagram illustrating a display position.

FIG. 13 is a diagram illustrating total data.

FIG. 14 is a diagram illustrating literal information.

FIG. 15 is a diagram illustrating a code.

FIG. 16 is a diagram illustrating an example of form data generation.

FIG. 17 is a diagram (continued) illustrating an example of form data generation.

FIG. 18 is a diagram showing an example of total data composed of multi-dimensional and multi-layers.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Display / operation part 2 Control part 3 Display data generation part 4 Display information access part 5 Display information storage part 6 DB access part 7 Multidimensional database 8 Total data storage part 9 Literal information storage part 10 Table of count origin 11 DB access part 12 Automatic literal table generation mechanism 13 Data reading unit 14 Matching processing unit 15 Tree structure creation unit 16 Literal table sorting unit 17 Literal table control unit 18 Data high-speed aggregation mechanism 19 Data reading unit 20 Total matrix creation unit 21 Corresponding cell・ Search unit 22 Data addition unit 23 Literal table control unit

Claims (2)

[Claims] 1. A literal which is written on a form, a code indicating a hierarchy in which the literal is written, a code indicating the number of the literal in the hierarchy, a parent code, and a literal having literal information indicating the number of children. A multidimensional analysis system comprising: a literal table storage unit that stores a table for each of a plurality of axes; a total data storage unit that stores total data written on a form; and a literal table generation mechanism. -The table generation mechanism reads one record from the table of the counting source, and checks whether the literal included in the read record exists in the literal table of the corresponding axis. Adds the literal information corresponding to the literal to the corresponding axis literal table, Correcting the literal information of the literal of the parent of the literal. 2. A literal which has a literal written on a form, a code indicating the hierarchy in which the literal is written, a code indicating the number of the literal in the hierarchy, a parent code, and literal information indicating the number of children. A multidimensional analysis system comprising: a literal table storage unit that stores a table for each of a plurality of axes; a total data storage unit that stores a total matrix in which total data written on a form is written; and a total data generation mechanism The summary data generation mechanism includes a process of generating a summary matrix in which a column corresponds to a literal on the vertical axis and a row corresponds to a literal on the horizontal axis, a process of reading one record from a table of the aggregation source, a literal -While referring to the table, correspond to the literal on the horizontal axis and the literal on the vertical axis specified by the read record. A multidimensional analysis system, comprising: performing a process of searching for a cell of an aggregation matrix; and a process of increasing a numerical value of a searched cell by a numerical value included in a read record.