JP4339558B2 - Multidimensional data processing method by computer - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method for multi-dimensionally processing data stored in a storage device with a computer, and in particular, a wide variety of data composed of three or more types of classification items and one or more types of numerical items is obtained. The present invention relates to a multi-dimensional data processing method by a computer used for managing, analyzing, or utilizing data according to the purpose and application by counting and searching as dimensional data.
[0002]
[Prior art]
When data stored in a computer storage device is aggregated as multidimensional aggregated data and this multidimensional aggregated data is to be searched and used, a multidimensional processing database called MDB (Multi-dimensional Data Base) is introduced. Or use the multi-dimensional processing function provided as a component, design and build a dedicated application system on the user side, and use that system to aggregate and accumulate data in multiple dimensions. In general, a search process is performed on the aggregated data.
[0003]
[Problems to be solved by the invention]
However, development of this kind of application system requires knowledge and technology of MDB or multidimensional processing functions, specifically, DB multi-dimension model, cube architecture in data warehouse and OLAP. It is necessary to make software design and program creation of the system using knowledge and technology such as definition and design, various objects, components, interfaces, customization development languages and tools, etc. There are few people or experienced people, and it is very difficult for users to develop themselves. In addition, the development of application systems using MDB is generally more difficult than application system development by the introduction of RDB (Relational Data Base), which is relatively widespread, and the burden is large. Takes time and money. Also, a considerable burden is required for maintenance after the start of operation.
[0004]
Here, if the multidimensional aggregated data is considered as a set of 2D aggregated data, the user can easily understand the 2D aggregated data. It can be used immediately as a view. Thus, in order to be able to directly search for arbitrary two-dimensional aggregated data from multidimensional aggregated data, if the user is aware of the two-dimensional aggregated view composed of rows and columns, the search condition Must be specified. In addition, the user can easily specify the search condition by selecting a button so that multi-dimensional aggregate data can be automatically segmented and the corresponding two-dimensional aggregate data can be retrieved simply by selecting an operation button. It is necessary to.
[0005]
The present invention has been made in view of the above circumstances, does not require knowledge or technology of MDB or multidimensional processing function, and can develop and use multidimensional aggregation of source data without developing a dedicated application system. A multi-dimensional data processing method by a computer that enables a user to easily search an aggregate view in the form of a two-dimensional aggregate data table that can be easily seen from multi-dimensional aggregate data by simply selecting a button. The purpose is to provide.
[0006]
[Means for Solving the Problems]
The invention described in claim 1 has three or more kinds of classification items and one or more kinds of numerical items, and records source data for multidimensional tabulation in a record format stored in a storage device. , A multi-dimensional tabulation method by a programmed computer, comprising: (i) a multi-dimensional tabulation key item comprising three tabulation key items: a field tabulation key item, a row tabulation key item, and a column tabulation key item; An area is stored in the storage device for storing a large number of multidimensional aggregated data in a record format composed of numeric aggregate items that aggregate the numeric data of the numeric items by the multidimensional aggregate key item and store the result. , (Ii) The classification item Of 3 More than ,in front Of the numeric value field Numeric Use it as a summary key item by dimension, which is the key when summarizing data by dimension. The One of the summary key fields by dimension 3 types , Field of view total key item, Above Line total key item and Above One of the column summary key fields Respectively All combinations of those correspondences For each pattern , Field of view total key item, Above Line total key item and Above Column summary key field The data of the multi-dimensional aggregation key item is made to correspond to the data value of the aggregation key item by dimension, and the numerical data of the numerical item is made to correspond to the numerical aggregation item, and the record format for each pattern Create a large number of multidimensional aggregated data, and (iii) record the multidimensional aggregated data in record format for each pattern, The multi-dimensional aggregate key item Sort as key Multidimensional aggregation For each unique data value of the key field number Value term Aggregate numeric data for eyes And (iv) storing the totaled result in the corresponding area in the storage device as multidimensional aggregated data in a record format after the aggregation. By, Multi-dimensional of the source data Counting process is 1 Times It is a computer-implemented multidimensional data processing method characterized by the above.
[0007]
In this way, by storing the numeric item data aggregated for each unique data value of the multi-dimensional aggregate key item in which the view key, row, and column aggregate key items are connected in order, the numerical summary item is stored in three or more dimensions. Can be managed as a two-dimensional total data table composed of rows and columns for each field of view, regardless of the number of dimensions.
[0008]
The invention according to claim 2 ,in front In multi-dimensional aggregated data, when using multiple aggregation methods for the same type of numeric data, , That 2. The computer-based multidimensional data processing method according to claim 1, wherein a plurality of numerical data tabulated based on each tabulation method are stored in a nested manner.
[0009]
The invention according to claim 3 Two tables: a field of view index consisting of three items for storage of the field of view name, summary view name, and multidimensional summary data, and a viewpoint index composed of two items of viewpoint name and storage destination of individual summary views Space for storing Storage device Secure in, A view-by-view summary view consisting of a two-dimensional summary data table composed of the row summary key items and the column summary key items of the multi-dimensional summary data, wherein the numerical summary items are arranged at positions where rows and columns intersect. ,in front Created from multi-dimensional aggregated data, based on the type of view aggregated by field of view ,in front Set a field of view with the field name as the name of the dimension key item corresponding to the field total key item, and the field name of the field , Collection Create a field of view index consisting of the total view name and the storage location of the multidimensional aggregated data. Secured Storage device In the area inside, In the total key item by dimension corresponding to the visual field total key item Store For each unique data value that is created, an individual aggregate view is formed, a viewpoint is set based on this, and a viewpoint index composed of the viewpoint name of the viewpoint and the storage location of the individual aggregate view is created. Secured Storage device In the area inside, Based on the visual field index, the viewpoint index, and the multidimensional aggregate data Sight Operation buttons that automatically generate fields, summary views, and viewpoints as categories, and display devices View, aggregate view, and individual view buttons on the same screen for each category in parallel 3. The multidimensional data processing method by a computer according to claim 1, wherein the display is performed.
[0010]
As a result, the user can easily specify the search conditions for multi-dimensional aggregate data using the field of view, viewpoint, etc. by being conscious of the row and column items of the aggregate view mainly consisting of a two-dimensional aggregate data table. It becomes like this.
[0011]
The invention according to claim 4 An area for storing a focus index composed of two items of a numerical aggregation item name and a focus name is secured in the storage device, In the source data for multidimensional aggregation, when using a plurality of aggregation methods for numeric data of the same numeric item having a plurality of numeric items, the numeric aggregation items of the multidimensional aggregation data for each aggregation method In relation to Rush Set the point Attached corresponding to the numerical item and the aggregation method Create a focus index that consists of numerical summary item names and focus names Secured Storage device In the area inside, Based on this focus index and the multidimensional aggregate data, Above Operation buttons with categories for numerical summation items and focus are automatically generated and displayed on the display device. Placed on the same screen as the operation screen with the field of view, total view, and viewpoint as the category. 4. The multidimensional data processing method by a computer according to claim 3, wherein the data is displayed.
[0012]
According to a fifth aspect of the present invention, the operation buttons are automatically generated and displayed on the display device. Sight Individual buttons for field and summary views Is selected by the input device, The storage device Stored in that area Referring to the field of view index, find the storage location of the relevant multidimensional aggregate data, and Sight Point individual buttons Is selected by the input device, The storage device Stored in that area Referring to the viewpoint index, the storage location of the corresponding individual aggregate view is obtained, and this individual aggregate view is , Multidimensional aggregate data In the storage device Take it out of the area , On display , View, summary view, viewpoint, output format consisting of row and column summary key items 5. The multidimensional data processing method by a computer according to claim 3 or 4, characterized by displaying.
[0013]
As a result, only the operation buttons displayed on the display device are selected using the input device, that is, without knowing the knowledge or technology of the MDB or multidimensional processing function, and without developing a dedicated application system. However, by referring to each index stored in the storage device, the multi-dimensional aggregate data can be automatically separated and the corresponding two-dimensional aggregate data can be retrieved, so that the user can easily search for the necessary aggregate view. It becomes like this.
[0014]
According to a sixth aspect of the present invention, the operation buttons are automatically generated and displayed on the display device. ,number Individual buttons for value summary items and focus Is selected by the input device, Multidimensional aggregate data The individual aggregate view retrieved from the area in the storage device storing Corresponding to the relevant numerical summary items and focus Number Value data Limited to The ,in front 6. The computer-implemented multidimensional data processing method according to claim 5, wherein the individual total view is arranged at a predetermined position and the individual total view is displayed on a display device.
[0015]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
A computer system for carrying out the present invention comprises a central processing unit 10, a storage unit 12, a display unit 14, and an input unit 16, as shown in FIG. The storage device 12 stores multidimensional aggregation source data 20, multidimensional aggregation data 30, a visual field index 40, a viewpoint index 50, and a focus index 60.
[0016]
The source data 20 for multidimensional tabulation used in the present invention is a record format as shown in FIG. 2 (a), and has three or more types of classification items (tabulation key items by dimension) 22, One or more numerical items 24 are included. A specific configuration example of the source data 20 is shown in FIG. This source data 20a has three kinds of classification items (total key items by dimension) 22a, 22b, and 22c, and two kinds of numerical items 24a and 24b, and all of the classification items 22a, 22b, and 22c are included. It becomes a total key item by dimension. Here, the tabulation key item by dimension is an item that becomes a key when tabulating the data of the numerical items 24a and 24b in multi-dimensions. In this example, the number of dimensions such as dimension 1, dimension 2, and dimension 3 is set. Grant and use.
[0017]
For example, the classification item 22a of the source data 20a shown in FIG. 2B includes a product name such as “PC”, the classification item 22b includes a store name such as “Shinjuku store”, and the classification item 22c includes Dates such as “April 2002” are stored respectively. Hereinafter, the dimension 1 aggregation key item (dimension aggregation key item 1) corresponding to this classification item 22a is referred to as “product”. The total key item of dimension 2 corresponding to item 22b (total key item 2 by dimension) is “store”, and the total key item of dimension 3 corresponding to classification item 22c (total key item 3 by dimension) is “year and month”. To do. Further, the numerical item 24a of the source data 20a shown in FIG. 2B stores the sales quantity such as “10”, and the numerical item 24b stores the sales amount such as “1500000”. Hereinafter, the numerical item 1 corresponding to the numerical item 24a is referred to as “sales quantity”, and the numerical item 2 corresponding to the numerical item 24b is referred to as “sales amount”. An example in which the following multidimensional aggregate data 30 is created based on the multidimensional aggregate source data 20a will be described.
[0018]
Note that each of the dimension total key items 22a, 22b, and 22c can have a hierarchical structure, but here it is assumed that there is no hierarchy for easy understanding. In other words, in the case of having a hierarchical structure, multi-stage data aggregation using a plurality of aggregation keys may be executed for each dimension, but the description thereof is omitted here.
[0019]
As shown in FIG. 3, the multi-dimensional aggregate data 30 includes a visual field total key item 32a, a row total key item 32b, and a column total key item 32c, and one or more types, in this example, two types of numerical total items 34a and 34b. It is data in record format composed of Regardless of the number of dimensions, a multi-dimensional tabulation key item 36 is formed by combining the field tabulation key item 32a, the row tabulation key item 32b, and the column tabulation key item 32c and connecting them in order. Each multidimensional aggregated data 30 is sorted in ascending or descending order by the data value of the multidimensional aggregated key item 36, and then each numeric data item 24a of the source data 20a is assigned to each unique data value of the multidimensional aggregated key item 36. 24b is counted and stored in the numerical tabulation items 34a and 34b, and the multidimensional tabulation process can be executed in one pass. In this way, when the multidimensional aggregate data 30 is created based on the multidimensional aggregate source data 20, the dimension-specific aggregate key items 22a and 22b of the multidimensional aggregate source data 20a shown in FIG. 22c and the field of view, row, and column total key items 32a, 32b, and 32c shown in FIG. This method will be described with reference to FIGS.
[0020]
FIG. 4 shows, in the case of three-dimensional aggregation, associating the aggregation key items 22a, 22b, and 22c for each dimension of the source data 20a for multidimensional aggregation with the aggregation key items 32a, 32b, and 32c for the field of view, row, and column. An example in which the data of the multidimensional tabulation key item 36 is configured is shown. Here, as described above, the names of the total key items 22a, 22b, and 22c of the dimension 1, dimension 2, and dimension 3 of the source data 20a are “product”, “store”, and “year / month”. Then, the data of the multidimensional tabulation key item 36 is constituted by all the combinations corresponding to any of the field of view, row and column tabulation key items 32a, 32b, and 32c of the multidimensional tabulation data 30. This type of combination is called a pattern.
[0021]
For example, in Pattern 1, “Product”, which is the total key item 22a of dimension 1 of the source data 20a, is stored in the field total key item 32a, and “Store”, which is the total key item 22b of dimension 2, is stored in the row total key item 32b. “Year and month”, which is the total key item 22c of dimension 3, is associated with the column total key item 32c. Similarly, the pattern 2 includes “Product”, “Year / Month”, and “Store” that are the key items 22a, 22c, and 22b of dimension 1, dimension 3, and dimension 2, respectively, in order of view, row, and column total key items. It corresponds to 32a, 32b, 32c. The same applies to the subsequent patterns. In this way, in the three-dimensional case, there are a total of six types of patterns (number of combinations of 1 × 2 × 3). FIG. 4 shows all the patterns. A pair of rows and columns is expressed in a row * column format, and a two-dimensional aggregate data table composed of the rows and columns is called an aggregate view.
[0022]
FIG. 5 shows an example in which the same association as in FIG. 4 is performed and the data of the multi-dimensional tabulation key item 36 is configured in the case of four-dimensional tabulation. In this example, the source data 20a shown in FIG. 2 (b) is added to another category item, that is, a category item in which the aggregation key item of dimension 4 (aggregation key item 4 by dimension) is “customer”. An example is shown. In this case, the total key items of dimension 1, dimension 2, dimension 3, and dimension 4 of the source data are set to “product”, “store”, “year / month”, and “customer”, respectively. The data of the multi-dimensional tabulation key item 36 is constituted by all the combinations corresponding to any of the row and column tabulation key items 32a, 32b, and 32c. There are a total of 24 patterns (number of combinations of 1 × 2 × 3 × 4). FIG. 5 shows patterns 1 to 6 (only the pattern in which “product” is associated with the visual field totaling key item 32a). The same applies to the other patterns 7 to 24 (patterns in which “store”, “year / month”, and “customer” are associated with the visual field total key item 32a, respectively). FIGS. 4 and 5 show examples of the data structure of the multidimensional aggregation key item 36 in the case of three-dimensional aggregation and four-dimensional aggregation, but the same applies to the case of more than five-dimensional aggregation.
[0023]
Next, an example will be described in which the numerical aggregation items 34a and 34b in the multidimensional aggregation data 30 shown in FIG. 3 are configured by numerical data aggregated using a plurality of aggregation methods. When a plurality of aggregation methods are used for the same kind of numerical data, the numerical aggregation items 34a and 34b store a plurality of numerical data aggregated based on the respective aggregation methods in a nested manner. Fig. 6 shows the case where the numerical items 24a shown in Fig. 2 (b) are tabulated using three types of tabulation methods: "Actual results for the previous year", "Actual results for the current year" and "YoY comparison". An example is shown. In addition, all are counted by month. The previous year results are the results of the previous year of sales volume, which are totaled for each unique data value of the multi-dimensional aggregation key item 36, and the current year results are the same as the current year results of sales volume. . The year-on-year comparison is the ratio obtained by dividing the total sales volume for the current year by the previous year.
[0024]
In this way, in the record format data composed of the three types of visual field, row and column total key items 32a, 32b and 32c and one or more types of numerical total items 34a and 34b stored in the storage device 12. The data of the numeric items 24a and 24b aggregated for each unique data value of the multi-dimensional summary key item 36 to which the visual key, row and column summary key items 32a, 32b and 32c are linked are stored in the numerical summary item 34a, By storing in 34b, it becomes possible to manage three-dimensional or more multidimensional aggregated data as a two-dimensional aggregated data table composed of rows and columns for each field of view, regardless of the number of dimensions.
[0025]
Next, the setting of the visual field for the total view of the two-dimensional total data table will be described by taking the case of three-dimensional total as an example. FIGS. 7A to 7C show the field of view and the total view in Pattern 1, Pattern 3, and Pattern 5 of FIG. In pattern 1 (see Fig. 7 (a)), the field of view is set to "product", and the total view 38a for each field of view is "store * year / month". Yes. Similarly, in Pattern 3 (see FIG. 7B), the field of view is set to “Store”, and the view-by-view aggregate view 38b is “Product * Year / Month”, with the product in the row and the date in the column. It corresponds. Furthermore, in Pattern 5 (see Fig. 7 (c)), the field of view is set to "Year / Month", and the aggregate view by field of view 38c is "Product * Store", with products corresponding to rows and stores corresponding to columns. In each pattern, “Sales volume” is included as a numerical summary item, and patterns 2, 4, and 6 are the ones in which the rows and columns of patterns 1, 3, and 5 are replaced, In this way, in the case of three-dimensional aggregation, there are three types (6 patterns / 2) of aggregate views composed of unique matrices, and each has a field of view. There are types (2 patterns / 2) of aggregate views, and in the case of 4D aggregation, the aggregate views increase to 12 types (24 patterns / 2), and one view (product) as shown in FIG. There are 3 types (6 patterns / 2) of total views And thus to be standing.
[0026]
FIG. 8 shows, in the case of three-dimensional totalization, a storage device 12 as a visual field index 40, which is a table composed of the visual field name, the total view name, and the storage destination of the multidimensional total data based on the set “field of view”. An example stored in is shown below. When the field of view name is “product”, the aggregation view is “store * year / month”, and the storage destination of the multidimensional aggregation data 30 is the storage address of the aggregation data based on the pattern 1 in FIG. That is, the total data is stored in the area of the multidimensional total data 30 corresponding to the corresponding visual field, row, and column. Similarly, when the visual field name is “store” or “year / month”, the storage destination of the multi-dimensional aggregated data 30 is the storage address of the aggregated data based on the pattern 3 or the pattern 5 in FIG.
[0027]
Although not shown, when the aggregation key items of dimensions 1 to 4 are “product”, “store”, “year / month”, and “customer” respectively, the field of view name and the aggregation view name (parentheses) Inside) is as follows. Since there are three types of aggregate views for each field of view, the index becomes large.
Product (store * date), product (store * customer), product (customer * date),
Store (product * date), store (product * customer), store (customer * year),
Date (product * store), date (product * customer), date (store * customer),
Customer (product * date), customer (product * store), customer (store * date)
[0028]
After setting the field of view, the view for each field of view totalized by the unique data value of the field totaling key item 32a is individualized, and the viewpoint is set based on this unique data value. explain. FIG. 9 is an example in which the viewpoint is set using the visual field in the pattern 1 of FIG. That is, in this example, the unique data value of “product” which is the visual field total key item 32a in the pattern 1, that is, the visual field-based total view is individualized for each product. A unique viewpoint that personalizes this view-by-view summary view is called a viewpoint. In FIG. 9, for example, a personal computer (product name) stored in the classification item 22a in FIG. A total view 54b that is individualized with the product name) as the viewpoint 52b and a total view 54c that is individualized with the camera (product name) as the viewpoint 52c are shown. The pattern 2 is obtained by replacing the rows and columns of the pattern 1 and has the same viewpoint. Although not shown in the figure, in the same way, for patterns 3 and 4, for each “store”, and for patterns 5 and 6, “year” and “month” are individualized aggregate views (product * year and product * store). ).
[0029]
FIG. 10 shows, in the case of three-dimensional aggregation, a table composed of viewpoint names and storage locations of individual aggregate views based on the set field of view and viewpoint as a viewpoint index 50 for each field of view in the storage device 12. A stored example is shown. FIG. 10A shows a viewpoint index 50a when the field of view is “product” (patterns 1 and 2 in FIG. 4). The viewpoint names are, for example, a personal computer, a television, and a camera. In the storage destination, the storage address of the total data of the personal computer, television, and camera is entered. Each aggregated data is stored in an area of the multidimensional aggregated data 30 in which the field of view corresponds to the product, the row corresponds to the store, and the column corresponds to the year and month. FIG. 10B shows the viewpoint index 50b when the field of view is “store” (patterns 3 and 4 in FIG. 4), and FIG. 10C shows the field of view “year and month” (pattern 5 in FIG. 4). The viewpoint index 50c in the case of 6) is shown.
[0030]
After setting the viewpoint, the numerical total items 34a and 34b are arranged at positions where the rows and columns of the individual total views 54a to 54c intersect. In this case, when a plurality of tabulation methods are used for the data of the same numerical item, the focus 62 is set corresponding to each tabulation method, and is constituted by the numerical tabulation item name and the focus name. An example in which the table is stored in the storage device 12 as the focus index 60 is shown in FIG. The focus 62 and the focus index 60 are related to the numerical tabulation items 34a and 34b regardless of the pattern and the number of dimensions. In this example, there are two numerical total item names corresponding to the numerical items 24a and 24b shown in FIG. 2B, the numerical value 1 is “sales quantity”, and the numerical value 2 is “sales amount”. In this example, there are three focal points 62 for the sales volume. The names of the focal points are “FY1” for focus 1, “Current result” for focus 2, “Yearly” for focus 3, and “Sales amount”. Similarly, there are three focal points 62, and the focal point names are “plan” for focus 1, “actual” for focus 2, and “plan ratio” for focus 3.
[0031]
Based on the field index 40, viewpoint index 50, focus index 60, and multidimensional tabulation data 30 stored in the storage device 12 in this way, as shown in FIG. An operation button 72 whose category is “Aggregate View”, an operation button 74 whose category is “Viewpoint”, an operation button 76 whose category is “Numeric Total Item”, and an operation button 78 whose category is “Focus” are automatically generated. And displayed on the display device 14 (in the case of three-dimensional aggregation). Here, in this example, the operation button 70 whose category is “field of view” is composed of three individual buttons 70a of products, stores and years, and the operation button 72 whose category is “total view” is a store. * Year, product * year, month, product * Store consists of three individual buttons 72a, and the operation button 74 whose category is "viewpoint" is a personal computer, TV, camera (in the case where the field of view is a product, etc. Are omitted). In addition, the operation button 76 having a category of “numerical data items” is composed of two individual buttons 76a of sales volume and sales amount, and the operation button 78 having a category of “focus” is the current year result, the previous year result, the previous year result. It consists of three individual buttons 78a for the ratio.
[0032]
Here, it is assumed that an individual button with a halftone dot is selected in FIG. Specifically, using the input device 16, after selecting an individual product button with the visual field operation button in FIG. 12, the viewpoint, the numerical totaling item, and the focus operation button are used for the personal computer, sales volume, current year's results, Suppose you select an individual button for the previous year's results. The individual buttons enclosed in parentheses () among the operation buttons of the total view indicate that they are not displayed after the individual button of the product is selected. FIG. 12 shows a case of three-dimensional aggregation, and since the field of view and the aggregation view are 1: 1, there is no need to select individual buttons for the aggregation view.
[0033]
As described above, if the operation buttons 70 to 78 are automatically generated and displayed on the display device 14, the user is mainly aware of the individual aggregate views 54 a to 54 c and uses the input device 16 to view the visual field. By simply selecting the individual buttons 70a to 78a of the operation buttons 70 to 78 for the aggregation view, viewpoint, numerical aggregation item, and focus, it becomes possible to easily specify the search conditions for the multidimensional aggregation data 30 regardless of the number of dimensions. .
[0034]
Here, a flow of processing relating to selection of the operation buttons 70 to 78 will be described. First, when the individual product button 70a is selected with the visual field operation button 70 in FIG. 12, since the visual field and the total view are 1: 1, only the store * year / month operation button 72a is redisplayed in the total view. These two are hidden. Then, with reference to the visual field index 40 shown in FIG. 8, the storage address of the total data based on the pattern 1 in FIG. 4 is obtained, and the multidimensional total data 30 of the visual field (product) is taken out. Next, when the individual button 74a of the personal computer is selected by the viewpoint operation button 74 of FIG. 12, the storage address of the total data (store * year / month) of the personal computer is obtained with reference to the viewpoint index 50 shown in FIG. Takes the view of the viewpoint (computer) (two-dimensional summary data table). Furthermore, when the individual sales quantity button 76a is selected by the operation button 76 of the numerical totaling item shown in FIG. 12, and the individual button 78a of the current year result and the previous year result is selected by the focus operation button 78, the corresponding individual total view is displayed. The data of the current sales volume and the previous year's sales data is extracted from the aggregated data record to be configured. FIG. 13 shows an example in which the individual total view composed of the retrieved data is displayed on the display device 14.
[0035]
FIG. 13 shows a total view of a two-dimensional total data table (store * year and month) in which the field of view is a product and the viewpoint is a personal computer. The rows and columns are composed of the stores Shinjuku store, Shibuya store, and Ikebukuro store, respectively, in April 2002, May 2002, and June 2002. The numerical summary item is sales volume, and the focus is on current year and previous year results.
[0036]
As described above, even if an MDB is installed or a multi-dimensional processing function is used and a dedicated application system is not designed and constructed, the user can select the operation button displayed on the display device and store it. Since the multi-dimensional aggregate data stored in the apparatus can be automatically cut out and taken out as two-dimensional aggregate data, a necessary aggregate view can be easily searched.
[0037]
【The invention's effect】
As described above, according to the present invention, three-dimensional or more multidimensional aggregated data can be obtained as if a set of two-dimensional aggregated data is obtained regardless of the number of dimensions by using the field, row, and column aggregation key items. To come to think. Therefore, the user can easily understand as a summary table of a matrix based on the visual field, and can be immediately handled as a summary view for display and viewing.
[0038]
Also, considering the two-dimensional aggregate view consisting of rows and columns, the field of view, the aggregate view, the viewpoint, and the numerical aggregate items and focus search conditions are specified as necessary. Any two-dimensional aggregated data can be directly searched from the multidimensional aggregated data.
[0039]
Furthermore, search conditions can be specified simply by selecting an operation button. Multidimensional aggregated data is automatically separated based on the search conditions, and the corresponding 2D aggregated data can be easily retrieved. Will be able to.
[0040]
As described above, knowledge and technology of MDB or multi-dimensional processing functions are not required, and there is no need to develop a dedicated application system. The user only needs to select a button from the multi-dimensional aggregate data. You can easily search the 2D tabulated view and use it easily.
[Brief description of the drawings]
FIG. 1 is a diagram showing an example of a computer system for carrying out the present invention.
FIG. 2A is a diagram illustrating a general format of source data for multidimensional tabulation, and FIG. 2B is a diagram illustrating a specific configuration example of source data.
FIG. 3 is a diagram showing a general format of multidimensional aggregated data (in the case of 3D aggregation).
FIG. 4 is a diagram illustrating an example of the data configuration of a multidimensional aggregation key item (in the case of three-dimensional aggregation).
FIG. 5 is a diagram illustrating an example of a data configuration of a multidimensional tabulation key item (in the case of four-dimensional tabulation, the field of view is only a product).
FIG. 6 is a diagram illustrating a configuration example in the case of using three tabulation methods for numerical tabulation items in multidimensional tabulation data.
FIG. 7 is a diagram illustrating an example of setting different visual fields (in the case of three-dimensional aggregation).
FIG. 8 is a diagram illustrating an example of a visual field index (in the case of three-dimensional aggregation).
FIG. 9 is a diagram illustrating an example of individualizing a total view by a total field key item (in the case of three-dimensional totalization).
FIG. 10 is a diagram illustrating examples of different viewpoint indexes (in the case of three-dimensional aggregation).
FIG. 11 is a diagram illustrating an example of a focus index.
FIG. 12 is a diagram showing an example (in the case of three-dimensional aggregation) in which automatically generated operation buttons are displayed on a display device.
FIG. 13 is a diagram showing an example (in the case of three-dimensional aggregation) in which an aggregation view is retrieved from multidimensional aggregation data and displayed.
[Explanation of symbols]
10 Central processing unit
12 Storage device
14 Display device
16 input devices
20, 20a Source data
22, 22a, 22b, 22c Classification item (dimensional total key item)
24, 24a, 24b Numeric items
30 Multidimensional data
32a, 32b, 32c Total key items (view, row, column)
34a, 34b Numerical tabulation items
36 Multidimensional key items
38a-38c Total view by field of view
40 Field Index
50, 50a, 50b, 50c Viewpoint index
52a-52c Individual viewpoints
54a-54c Individual summary view
60 focus index
62 Focus
70-78 operation buttons
70a-78a Individual button

Claims (6)

This is a method for performing multidimensional aggregation processing by a programmed computer on record-type multidimensional aggregation source data having three or more types of classification items and one or more types of numerical items and stored in a storage device. And
(I) A multi-dimensional total key item composed of three total key items, a visual field total key item, a row total key item, and a column total key item, and the numerical data of the numeric item is totaled by the multi-dimensional total key item. An area for storing a large number of multi-dimensional aggregate data in record format composed of numerical summary items to store the results is secured in the storage device,
(Ii) three or more kinds of said classification items, using the numerical data of the previous SL numeric fields as dimensions totals key field comprising a key when aggregating treatment by dimension, or three of the dimensions totals key field and the field of view aggregation key fields, the row aggregation key fields and respectively corresponding to one of said column aggregation key fields, each pattern of all combinations of their corresponding said field aggregation key fields, the row aggregation key fields and creating a data multidimensional aggregation key item that associates the data values of the dimension-specific aggregation key fields in the column aggregation key fields, further, with the numerical data of the numeric fields to correspond to the numerical summary fields, the Create a large number of multi-dimensional aggregate data in record format for each pattern,
(Iii) multidimensional aggregated data record format for each of the patterns, the multi-dimensional aggregation key item sorted as a key, the number Neko th numerical data for each unique data value of the multi-dimensional aggregation key fields aggregates,
(Iv) By storing the aggregation results as multidimensional aggregated data in a record format after aggregation in the corresponding area in the storage device ,
A multidimensional data processing method by a computer, wherein the multidimensional aggregation processing of the source data is executed at once . Prior Symbol multidimensional aggregate data, when using a plurality of calculation method for the same type of the numerical data, the said numerical summary fields, a plurality of numerical data aggregated based on the aggregation method of their respective 2. The computer multidimensional data processing method according to claim 1, wherein the data is stored in a nested manner. Two tables: a field of view index consisting of three items for storage of the field of view name, summary view name and multidimensional summary data, and a view index composed of two items of viewpoint name and storage destination of individual summary views Secure an area in the storage device to store
A view-by-view summary view consisting of a two-dimensional summary data table composed of the row summary key item and the column summary key item of the multidimensional summary data, and the numerical summary item is arranged at a position where a row and a column intersect . created from the previous Symbol multidimensional aggregate data,
Based on the type of the field of view by aggregate view, the previous Symbol name of the dimension-specific aggregation key item corresponding to the field of view aggregation key item to set the field of view to a field name, the field name of the visual field, aggregated view name, multi Create a field of view index configured by the storage location of the dimension summary data, store it in the area in the secured storage device ,
A unique tabulation view is formed for each unique data value stored in the tabulation key item for each dimension corresponding to the visual field tabulation key item, and a viewpoint is set based on this view. Create a viewpoint index configured by the storage location of the aggregate view, and store it in the reserved storage area.
The field of view index, based on the viewpoint index and the multidimensional aggregate data, visual field, automatically generates an operation button to an aggregated view and perspective category field on the same screen of the display device, aggregate views and perspectives The multi-dimensional data processing method by a computer according to claim 1 or 2 , wherein the individual buttons are collectively displayed in categories and displayed in parallel . An area for storing a focus index composed of two items of a numerical aggregation item name and a focus name is secured in the storage device,
In the source data for multidimensional aggregation, when using a plurality of aggregation methods for numeric data of the same numeric item having a plurality of numeric items, the numeric aggregation items of the multidimensional aggregation data for each aggregation method in association with, and set the focus point, the numeric fields and numerical summary field name attached in correspondence with the aggregation method, to create a focus index constituted by the focal name, region in the secured storage device Stored in
Based on the focus index and the multidimensional aggregate data, the numerical summary fields and the focus automatically generates an operation button to categories, a display unit, an operation button and the field of view, category summary view and the view 4. The multidimensional data processing method by a computer according to claim 3, wherein the multidimensional data processing method is arranged and displayed on the same screen as the displayed screen . Among the operation buttons that are automatically generated and displayed on the display device, viewed when selected by the input device an individual button of the field and aggregated view, with reference to the field of view index that is stored in the area in the storage device, Find the storage location of the relevant multidimensional aggregate data,
Further, when selected by the input device an individual button of the visual point, with reference to the viewpoint index that is stored in the area in the storage device, asking the storage destination of the corresponding said discrete aggregate view,
The individual aggregate view, removed from the area in the storage device that stores the multidimensional aggregate data,
5. The computer-based multidimensional data processing method according to claim 3 or 4, wherein the data is displayed on a display device in an output format comprising a field of view, a total view, a viewpoint, and a total key item of rows and columns . Among the operation buttons that are automatically generated and displayed on the display device, when selected by the input device a separate button numeric summary fields and focus, removed from the area in the storage device that stores the multidimensional aggregate data said limited to numerical data that corresponds to the relevant numerical summary field and focus of the individual in the aggregate view was, prior SL is arranged at a predetermined position of the individual aggregate view, on a display device the individual separate aggregate view 6. The multidimensional data processing method by a computer according to claim 5, wherein the display is performed.

Images