JPH11184843A - Data sum-up device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device which can execute a multidimensional sum-up operation processing with simple hardware constitution. SOLUTION: The data sum-up device 1 executes the multidimensional sum-up operation by using plural previously sorted records having n-pieces of item data and corresponding data of a sum-up object. The data sum-up device 1 is provided with (n+1)-pieces of data sum-up darts 2-4 having key break detection part detecting key breaks different in item data of mutual keys when plural items designated in higher orders are set to be keys against two continuously inputted records, and sum-up operation parts 10 summing up data of the sum-up object of the record, whose item data of the keys are similar, and generating sum-up data and outputting the record where sum-up data when the key break is detected by the break detection part is made to correspond to item data of the key corresponding to the data in series in order from the largest number of the items of the key by varying the number of the items of the respective keys.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data summarizing apparatus for performing a summation operation using a summation target data of a plurality of records including a plurality of item data and a corresponding summation target data. The present invention relates to a data tabulation device that performs a tabulation operation using data to be tabulated for each sorted record.

[0002]

2. Description of the Related Art In general, a tabulation calculation process is performed as a process performed using a database in which a plurality of records each including n (n is an integer) item data and corresponding data to be tabulated are compiled. . In this tabulation calculation process, tabulation calculation is performed using tabulation target data corresponding to the same item data, and tabulation data is generated for each item data.

At this time, a tallying operation device for performing tallying operation includes:
When records having the same item data are not continuously input, the tabulation operation device must sort the input records for each item data and then perform the tabulation operation, or perform the tabulation operation while sorting. . For this reason, the load on the tally calculation device is large, and the tally calculation processing is slowed down.

Therefore, using a sorter (a hardware sorter called a sort processor), the records are sorted (rearranged) in advance in the database in the order of the higher order items. Thereby, since the records can be input to the totaling device in the sort order, the totalizing device does not need to sort each input record. As a result, the device that performs the tallying calculation can quickly perform the tallying calculation.

[0005] Aggregation calculation processing (roll-up processing) includes one-dimensional aggregation calculation processing and multidimensional aggregation calculation processing. The one-dimensional tabulation calculation processing refers to generating tabulation data for each of the n item data by performing a tabulation calculation using data to be tabulated (for example, sales data) corresponding to the n item data. . Since the processing of the one-dimensional tabulation operation is relatively simple, dedicated hardware for the one-dimensional tabulation operation has also been created. However, 1
Only total data for each of n item data can be generated from one database.

[0006] Therefore, there is a multidimensional totaling calculation process as a process capable of obtaining a large amount of information (total data) from one database.

[0007] The multi-dimensional tabulation operation process is to use the tabulation target data corresponding to the n item data, and for each of the n item data specified in the higher order, n-1 specified in the higher order.
, For each of the n-2 item data specified in the higher order,. . . . . This means that total data is generated for each item data specified in the order of higher rank and for each item data of 0 items.

[0010] The generation of the total data for each of the zero item data means that the total data is generated by performing the total operation on the total data of all the input records.
As described above, in the multidimensional totaling calculation process, a large amount of information can be obtained from one database. Note that this multi-dimensional total calculation processing is performed by using a database such as a relational database or the like to execute a sequence of instructions for the multi-dimensional total calculation processing by the CPU.

A conventional technique for performing a multidimensional totaling calculation process using each sorted record will be described below. FIG. 11 is a configuration diagram showing a system (hereinafter, simply referred to as a system) that performs a multidimensional totaling calculation process.

(System Configuration) The system includes a sorting device 80 for sorting a plurality of records in the order of higher-order items in a database composed of a plurality of records consisting of n items of data and corresponding data to be counted. When,
Based on the above-described total calculation processing instruction sequence, using a storage unit 81 in which a total calculation processing instruction sequence for instructing the multidimensional total calculation processing is stored, and totaling target data of each record output from the sorting device 80 in the sort order. , CPU that performs aggregation calculation
82. The CPU 82 has a plurality of tallying operation counters for tallying the tallying target data of each record. These tally calculation counters store tally data calculated tally each time each record is input, for each input of each record. For example, in each tally counter, tally data, which is the result of tally calculation each time a record is input, is stored at each predetermined position. Also, the CPU 82
Has a register for storing each record generated by the multidimensional aggregation operation.

(Operation of System) The operation of this system will be described below. First, the sorting device 80 sorts a plurality of records of the database to be subjected to the tabulation operation in the order of the higher-ranked items. FIG. 12 is a diagram showing the database 92 sorted. Each record (record 92
a to record 92q) are item data (Fujisawa, Kamakura) for item 1 (city name), item data (TV, radio, etc.) for item 2 (product name), and tabulation target for item 3 (tabulation item) Data (sales data)
Consists of Then, the database is sorted as follows. That is, each record in the database is
The item data of item 1 of each record is sorted in the order of Fujisawa and Kamakura. Then, without breaking the sort order of the item data of item 1 (upper item), each record in the database is sorted with respect to the item data of item 2 of each record in the order of television, video, and stereo. You.

Second, the CPU 82 reads out and decodes a tallying operation instruction sequence from the storage unit 81 and sets so that the tallying operation can be performed.

(Setting operation of total calculation processing) This setting operation will be described below. The CPU 82 is set to perform a three-dimensional totaling operation. That is, the CPU 82 performs a tallying operation of the tallying data corresponding to the item data of the item 1 and the item 2, a tallying operation of the tallying data corresponding to the item data of the item 1, a tallying operation of all the tallying data, Is set to be able to be performed.

At this time, in order for the CPU 82 to perform the counting operation, it is necessary to generate the counting data by performing the counting operation for each item data of the item 1 and the item 2 (and the item 1). For this reason, the CPU 82 needs to perform the tallying operation of the tallying target data corresponding to the same item data, detect the change of the input item data, and newly perform the tallying operation again. As a result, the CPU 82
Then, the following settings are performed.

That is, the CPU 82 uses the item 1 and the item 2 as a first key to detect a first key break in which the item data of the first key is different from each other in two records input in the sort order. Is set. Also, CPU
The item 82 is set so that item 1 is used as a second key, and a second key break in which the item data of the second key is different from each other is detected for the two records. And CP
The first tabulation calculation counter of U82 is set so as to tabulate the tabulation target data of records having the same item data of the first key. Similarly, the second counting operation counter of the CPU 82 is set so as to count the counting target data of the records having the same item data of the second key. It should be noted that the first tally calculation counter and the second tally calculation counter are set so that when the first key break and the second key break are detected, respectively, the tally calculation operation is reset and a new tally calculation operation is performed. Is done. Then, the third counting operation counter of the CPU 82 is set so as to count all the counting target data of each input record.

Third, the CPU 82 instructs the sorting device 80 to output each record. As a result, the sorting device 80 stores each record in the CPU 8 in the sort order.
Output to 2.

Fourth, the CPU 82 performs a multidimensional totaling operation on the records input in the sort order. FIG. 13A is a diagram showing each tabulated data value stored in each predetermined position of each tabulation operation counter, and FIG.
When each record of FIG. 12 is input to the CPU 82,
FIG. 9 is a diagram showing a tabulation calculation result database in which records output from a CPU 82 are put together.

(Operation of CPU when Each Record is Input) The operation of the multidimensional totaling calculation process by the CPU 82 will be described with reference to the case where each record of FIG. 12 is input as an example. It should be noted that, among the input records 92a to 92q, (1) when the calculation operation is started (when the record 92c is input from the record 92a), (2) when the first key break is detected ( The case where the record 92d is input will be described as an example),
(3) When the first key break and the second key break are detected (the case where the record 92i is input will be described as an example), and (4) When the tallying calculation process ends (the record 92q is input). ) Will be described.

(1) Records 92a to 92c
Is input, the item data of the first key and the second
The item data of the key is the same as each other. Therefore, the first
Since the key break and the second key break are not detected, the tallying calculation counters 93, 94, and 95 calculate the tallying target data. At this time, record 92
When "a" is input, each of the tally calculation counters 93, 94, 9
In each of the predetermined positions 93a, 94a, and 95a of No. 5, the tabulation target data value 8 of the record 92a is stored as tabulation data. When the record 92b is input, each predetermined position 93b, 94b, 9 of each of the counting operation counters 93, 94, 95
5b stores a tabulated data value 18 which is a result of tabulating calculation (addition) of the tabulated data value 10 of the record 92b to the tabulated data value 8 described above. And record 9
When 2c is input, each of the tally calculation counters 93, 94,
In each of the predetermined positions 93c, 94c, and 95c, a totalized data value 2 which is a result of performing a totalization operation (addition) of the totaled data value 2 of the record 92c to the totalized data value 18 is provided.
0 is stored.

(2) When the record 92d is input to the CPU 82, the item data of the first key of the previously input record 92c is Fujisawa (item data of item 1) and television (item data of item 2). Of record 92d
Since the item data of the first key is Fujisawa and video, the first key break is detected. Then, in the CPU 82, the record 97a (FIG. 13 (b)) in which the total data value 20 stored at the predetermined position 93c of the first total calculation counter 93 is associated with the item data (Fujisawa, TV) of the first key. ) Is stored in the register. Then, the counting operation of the first counting counter 93 is reset, and the counting target data value 9 of the record 92d is stored in the predetermined position 93d of the first counting counter 93. Then, the first counting operation counter 93 newly performs a counting operation.

At this time, since the second key break is not detected, the predetermined position 94d of the second total calculation counter 94 stores the total data value 20 (total data until the record 92c is input) and the record 92d of the record 92d. A tabulated data value 29 obtained by tabulating (adding) the tabulated data value 9 is stored. Predetermined position 95 of third total calculation counter 95
In d, a total data value 2 obtained by performing a total operation (addition) of the total data value 9 (total data up to the time when the record 92c is input) and the total data value 9 of the record 92d.
9 is stored.

(3) When the record 92i is input, the item data of the first key of the previously input record 92h is Fujisawa and stereo, whereas the item data of the first key of the record 92i is Kamakura. , A first key break is detected. Then, the CPU 82 calculates the totaled data value 6 (the result of totalizing the target data of the record in which the item data of the first key is Fujisawa and stereo) stored in the predetermined position 93 e of the first totaling operation counter 93, and One key item data (Fujisawa, stereo)
Is stored in the register. Then, the counting operation of the first counting counter 93 is reset, and the counting target data value 2 of the record 92i is stored in the predetermined position 93f of the first counting counter 93.

The item data of the second key of the previously input record 92h is Fujisawa, while the item data of the second key of the record 92i is Kamakura.
Key break is detected. Then, in the CPU 82,
The total data value 38 (result of total operation of the total data of the record whose key item data is Fujisawa) stored at the predetermined position 94e of the second total operation counter 94 and the second
A record 97c in which key item data (Fujisawa, ALL) is associated with is stored in a register. At this time, the item data of item 2 is given data ALL indicating that all the item data of item 2 is included. Then, the tallying operation of the second tallying calculation counter 94 is reset, and the tallying target data value 2 of the record 92i is changed to the second tallying value.
It is stored at a predetermined position 94f of the tallying operation counter 94.
In the third tabulation calculation counter 95, the tabulation data value 40 obtained by tabulation calculation (addition) of the tabulation target data value 2 of the record 92i is added to the tabulation data value 38 (tabulation data until the record h is input) at a predetermined position. 95f.

(4) When the last record 92q is input, the tally data value 16 of the first tally calculation counter 93 (the result of tallying the tally data of the record whose key item data is Kamakura and stereo) and , A record 97d associated with the first key item data (Kamakura, stereo) is stored in the register. Similarly, the aggregated data value 60 of the second aggregated operation counter 94 (the result of the aggregation operation of the aggregated data of the record whose key item data is Kamakura) and the second key item data (Kamakura, ALL) are The associated record 97e is stored in the register. Then, a record 97f in which the total data value 98 of the third total calculation counter 95 (total data until the record 92q is input) and the item data (ALL, ALL) are stored in the register. . At this time, ALL indicating that the item data of items 1 and 2 of the record 97f includes all the item data of items 1 and 2
Is given.

All the records (records 92a to 92q) sorted in this way are stored in the CPU 82.
When the key break detection processing and the tabulation calculation processing are completed, the following processing is performed. That is, C
In the PU 82, each record in which each item data of the first key is associated with each aggregation data, each record in which each item data of the second key is associated with each aggregation data, The record including the aggregation data as a result of the aggregation operation of the target data and the record including the aggregation data are read from the register. Then, the CPU sorts each of the records in the order of the higher-ranking items, and outputs a totaling result database 97 shown in FIG. 13B.

[0026]

However, in the above-mentioned prior art, one CPU 82 executes a sequence of instructions for instructing multi-dimensional total calculation at one time.
There is a problem that an excessive load is applied to the CPU 82.
For this reason, the tallying calculation process becomes slow and the CPU 82
However, there is a problem that other processings to be performed are also delayed.

On the other hand, conventionally, when an apparatus for performing a multidimensional aggregation process is to be realized by dedicated hardware, the configuration becomes complicated and it is difficult to realize the apparatus.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and a first object of the present invention is to provide an apparatus capable of performing a multidimensional totaling calculation process with a simple hardware configuration. It is in.

A second object of the present invention is to provide an apparatus for performing a multi-dimensional totaling operation using a CPU, which is high-speed processing and does not apply an excessive load to the CPU.

[0030]

In order to achieve the above object, a data tallying apparatus according to a first aspect of the present invention comprises n (n is an integer) item data and tallying data corresponding thereto. Are sorted in the order of the top items, and the data summarizing device performs the summing operation using the sorted data of each record. Key break detection means for detecting key breaks with different key item data when a plurality of selected items are used as keys, and tabulating and calculating tabulation target data of records having the same key item data Aggregation function for generating data and, when a key break is detected, outputting a record in which the aggregation data is associated with item data of a key corresponding to the aggregation data. Means, (n + 1) pieces of data collecting unit having found by varying the number of items each key one another, in the order item number of keys is large, those which are connected in series.

In the data totaling apparatus according to the second aspect of the present invention, a plurality of records each including n (n is an integer) item data and the corresponding data to be totaled are sorted in the order of the upper item, and each of the sorted records is sorted. In a data aggregation device that performs an aggregation operation using data to be aggregated in a record, when two or more records sequentially input in the sort order are used as a key for a plurality of items specified in an upper order, a mutual key is used. Key break detection means for detecting a key break in which the item data of the key is different, and calculating tabulation data on the tabulation target data of the record having the same key item data to generate tabulated data, and when a key break is detected, And a totaling operation means for outputting a record in which the totalized data is associated with the item data of the key corresponding thereto.
1) Data totaling means are connected in parallel with the number of items of each key being different from each other.

[0032] The data tallying apparatus according to the third invention is characterized in that
Alternatively, in the second invention, there is provided an output control unit that sorts the records output from the tally calculation unit of each of the data tally units and sorts the records in the order of the higher order items, and puts them together in the output order.

According to a fourth aspect of the present invention, in the data totaling apparatus, a plurality of records each including n (n is an integer) item data and the corresponding data to be totaled are sorted in the order of the upper item, and each of the sorted items is sorted. In a data aggregation device that performs an aggregation operation using data to be aggregated in a record, when two or more records that are sequentially input are used as a key for a plurality of items specified in a high-order order, items of a mutual key are used. A key break detection step of detecting key breaks in which data is different, and a tabulation operation of tabulation target data of records having the same key item data to generate tabulation data; and, when a key break is detected, the tabulation data. And a counting operation step of outputting a record in which the item data of the key corresponding thereto are associated with each other.
1) storage means for storing data aggregation instructions,
And (n + 1) CPUs for decoding and executing (n + 1) data totaling instructions, respectively.
The number of key items is different from each other in the data totaling instruction sequence, and the (n + 1) CPUs are connected in series in descending order of the number of key items in the data totaling instruction to be decoded and executed. Things.

According to a fifth aspect of the present invention, in the data totaling apparatus, a plurality of records including n (n is an integer) item data and the corresponding data to be totaled are sorted in the order of the higher-ranked items, and each of the sorted records is sorted. In a data aggregation device that performs an aggregation operation using data to be aggregated in a record, when two or more records sequentially input in the sort order are used as a key for a plurality of items specified in an upper order, a mutual key is used. A key break detection step of detecting a different key break in which the item data of the key data is different, and a tabulation operation is performed on tabulation target data of records having the same key item data to generate tabulated data, and when a key break is detected, A (n + 1) number of calculation operation steps for outputting a record in which the aggregation data is associated with the item data of the key corresponding thereto. And (n + 1) CPUs for decoding and executing the (n + 1) data totaling instructions, respectively, wherein the (n + 1) data totaling instructions are: The number of key items is different from each other, and the (n + 1) CPUs are connected in parallel.

According to a sixth aspect of the present invention, there is provided a data totaling apparatus comprising:
Alternatively, in the fifth invention, the apparatus further comprises an output control means for sorting the records output from the CPUs in the order of output while sorting the records in the order of higher rank items.

[0036]

Preferred embodiments of the present invention will be described below with reference to the drawings.

Embodiment 1 The data aggregation device according to the first embodiment performs a multidimensional aggregation operation using a plurality of records including n (n is an integer) item data and corresponding aggregation target data. At this time, the plurality of records are sorted in advance by the sorting device in the order of higher-ranked items. In the present embodiment, as an example of the multidimensional totaling operation, two item data (item data of item 1 and item data of item 2) and corresponding data to be totaled (totaling data of item 3) A case will be described in which a three-dimensional totaling operation is performed using a plurality of records composed of.

(1) Configuration of Data Aggregation Device (1-1) Overall Configuration FIG. 1 is a block diagram showing the overall configuration of the data aggregation device according to the first embodiment. The data tallying apparatus 1 according to the first embodiment includes a plurality (three) of data tallying units 2, 3, and 4 for performing a plurality of (three) one-dimensional tallying operations, respectively, and each of the data tallying units 2, 3, And an output control unit 5 that receives each total calculation result record output from the control unit 4 and performs predetermined control and outputs the result.

In general, when performing a multidimensional totaling operation using a plurality of records composed of n item data and corresponding data to be totaled, the number of the data totaling units is set to (n + 1). ). The reason will be described below.

In the multidimensional totaling calculation process performed using the totaling target data corresponding to n item data, (n +
1) A dimensional aggregation calculation process is performed. That is, this (n +
1) In the dimensional tabulation calculation process, for each of n item data items specified in the higher order, for each of n-1 item data items specified in the upper order,. . . . , Total data is generated for each of the one item data designated in the higher order and for each of the zero item data. For this reason, it is necessary to prepare (n + 1) data tabulation units that respectively perform (n + 1) kinds of tabulation calculation processes.

(1-2) Data Tabulation Unit FIG. 2 is a diagram showing the configuration of the data tabulation unit 2. The configuration of the data tabulation units 3 and 4 is the same as the configuration shown in FIG. The data summarizing unit 2 detects a key break when two or more consecutively input records have a plurality of items specified in a higher order as a key, and detects a key break in which item data of each key is different. It has a part 6. The key break detection unit 6 stores continuously input records, and outputs a stored record when a new record is input, and a first record register 7 that outputs the stored record. A second record register 8 for storing each record, a key item data of each record newly stored in the first record register 7, and a key item of each record newly stored in the second record register 8. And a break detecting unit 9 for comparing the data with the data to detect a key break in which the item data of the keys are different from each other.

The data tabulation unit 2 tabulates the tabulation target data of the records having the same key item data to generate tabulation data, and when the break detection unit 9 detects a key break, tabulation. There is a counting operation unit 10 that outputs a record (total calculation result record) in which data is associated with item data of a key corresponding to the data.

Further, each data totaling section 2 has a record configuration information register 11 for storing the configuration information (items of the data to be totaled) of each record to be totalized according to an external instruction, and an external instruction. And a total calculation method instruction register 12 for storing information indicating that the total calculation method is specified.

In the configuration of each data tabulation unit of the present embodiment, each break detection unit 9 for detecting each key break and each tabulation calculation unit 1 for performing each tabulation calculation are provided.
The configuration is not limited to the above configuration as long as the configuration includes 0 and a specification unit that specifies an item of each key and an item of each data to be totaled.

The three data tabulation units configured in this manner are connected in series with the number of key items being different from each other and in descending order of the number of key items.

In the present embodiment, the number of data tabulation units arranged in the data tabulation device is not limited to three, but any number of data tabulation units can be arranged. For example, when performing a multidimensional totaling operation using a plurality of records including n items of data and corresponding target data, an (n + 1) -dimensional totaling operation is performed. In this case, (n +
1) The data totaling units are connected in series in a descending order of the number of key items, with different numbers of key items.

(1-3) Output Control Unit The output control unit 5 sorts the records (total calculation result records) output from the tally calculation unit 10 of each data tally unit in the output order while sorting them in the order of the higher-order items.

(2) Operation of Data Aggregation Device 1 The operation of the data aggregation device 1 configured as described above will be described below. Regarding this operation, (2-1) an operation for storing necessary information in the record configuration information register 11 and the tally calculation method instruction register 12 in response to an external instruction, and (2-2) an operation based on the information stored in each register. (2-3) Continuously inputting sorted records to the data tabulation unit 2 and causing each of the data tabulation units 2, 3, and 4 to output a tabulation calculation result record. Will be explained.

(2-1) Storage operation (2-1-A) Storage operation of each information in the record configuration information register 11 For example, an external device is connected to each of the data totaling devices 2, 3, and 4, By inputting each information from an external device, predetermined information is stored in the record configuration information register 11 of each of the data totaling devices 2, 3, and 4. This storage operation will be described in detail below.

By performing a predetermined operation using an external device, a target item of a key, a target item of a mask key, a mask data, and a target data item are specified. The specification of the target item of the mask key is to specify a target item for masking predetermined mask data (for example, ALL described later). When there is no item to be specified or when there is no data to be specified, information corresponding to a NULL value indicating that there is no item or data to be specified is specified.

The following information is stored in the record configuration information register 11 of each of the data totaling units 2, 3, and 4 by performing the above-described designations using an external device.

In the record configuration information register 11 of the data tabulation unit 2, key information indicating that items 1 and 2 are keys, tabulation target information indicating that item 3 is a tabulation target, and mask key target items are Mask key information indicating that the key does not exist is stored.

The record configuration information register 11 of the data totaling unit 3 stores key information indicating that item 1 is a key,
Mask key information indicating that item 2 is a mask key,
Counting target information indicating that item 3 is a counting target is stored.

The record configuration information register 11 of the data totaling unit 4 stores mask key information indicating that the items 1 and 2 are mask keys, totaling target information indicating that the item 3 is a totaling target, and a key target. Key information indicating that the item does not exist is stored.

(2-1-B) Operation of Storing Each Information in Total Calculation Method Instruction Register 12 Similarly, by performing a predetermined operation using an external device, the totalization of each data totaling device 2, 3, 4 is performed. The arithmetic method instruction register 12 stores predetermined information. This storage operation will be described in detail below.

By performing a predetermined operation using an external device, the information indicating one of the total calculation, the maximum calculation, the minimum calculation, the average calculation, etc., is calculated. It is stored in each tally calculation method instruction register 12. In the present embodiment, information for instructing the sum calculation operation is stored in each of the tally calculation method instruction registers 12 of the data tallying units 2, 3, and 4.

(2-2) Setting Operation After the above-described storage operation in each record configuration information register 11 and each totaling calculation method instruction register 12 is completed, each data totaling unit 2, 3, 4 sets each record configuration information register. The key information stored in 11 is sent to the break detector 9. As a result, the break detection unit 9 of each of the data tabulation units 2, 3, and 4 is set to detect a key break based on the key information. In addition, each data tabulation unit 2,
The aggregation operation units 3 and 4 perform the aggregation operation based on the aggregation target information read from each record configuration information register 11 and the information read from each aggregation operation method instruction register 12. Is set to

(2-3) Operation of inputting each record continuously in the sorted order to the data totaling unit 2 and outputting the totaling operation result records from each of the data totaling units 2, 3, and 4 Tabulated data for each completed record
The operation performed when the input is continuously input to is described. At this time, the operation of the data tabulation unit 2, the operation of the data tabulation unit 3, the operation of the data tabulation unit 4, the operation of the output control unit 5,
Will be described separately.

Note that, among the records in the database 92 of FIG. 12, the data is input to the data totaling unit 2 in order from the record 92a.

(2-3-1) Operation of Data Tabulation Unit 2 When the record 92a is input, the record 92a is first stored in the first record register 7. And
When the next record 92b is input, the first record register 7 transfers the stored record 92a to the second record register 8. Then, the transferred record 92a
Are stored in the second record register 8. Hereinafter, when a new record (92c, 92d ...) is input,
The above operation is performed between the record registers 7 and 8.

Then, based on the key information read from the record configuration information register 11, the break detection unit 9 sets the key of the record newly stored in the first record register 7 using the items 1 and 2 as keys. Is compared with the item data of the key of the record newly stored in the second record register 8. As a result of the comparison by the break detection unit 9, if the item data of the key is the same, the break detection unit 9 does not perform any operation. Then, the tallying calculation unit 10 performs the following tallying calculation using the tallying target data of the records stored in the second record register 8. That is, based on the information read from the tally calculation method instruction register 12 and the tally information read out from the record configuration information register 11, the tally calculation unit 10 calculates the tally target of records having the same key item data. An aggregation operation (total value calculation operation) is performed using the data (totaling data of item 3). This counting operation unit 10
Is continued until a key break signal described later is sent.

Further, as a result of the comparison by the break detecting section 9, when the key item data are different from each other, the break detecting section 9 detects a key break. For example, when the record 92d shown in FIG. 12 is input to the data totaling unit 2, the item data of the keys of the records stored in the record registers 7 and 8 are different from each other (the item data newly stored in the first record register 7). The key item data of the record 12d is Fujisawa and video, and the key item data of the record 12c newly stored in the second record register 8 is Fujisawa and television. For this reason, the break detection unit 9 detects a key break. Then, the break detection unit 9 sends the key break signal to the tally calculation unit 10.

When receiving the key break signal, the counting operation unit 10 counts the totaling data value 20 obtained by calculating the counting target data from the records 12a to 12c having the same key item data (Fujisawa, TV), and the key data. And outputs a total calculation result record in which item data (Fujisawa, TV) are associated with the item data. When a plurality of sorted records shown in FIG. 12 are input to the data totaling unit 2 in this manner, a database in which the totaling operation result records (records 21a to 21f) output from the data totaling unit 2 are collected. 21 is shown in FIG. Each tally calculation result record is sent to the output control unit 5 and also to the data tally unit 3. Then, the counting operation unit 10 resets the counting operation, and performs the counting operation using the counting target data of the record having the same item data as the key item data (Fujisawa, video) after the key break is detected. Do.

Then, for each input record,
As described above, each operation by each unit is repeated.

(2-3-2) Operation of Data Tabulation Unit 3 Each tabulation calculation result record (records 21a to 21 in FIG. 3A) output by the tabulation calculation unit 10 of the data tabulation unit 2.
1f) is input to the first record register 7 of the data totaling unit 3. The functions of the first record register 7 and the second record register 8 in the data tabulation unit 3 are the same as the functions of the record register of the data tabulation unit 2 described above.

The break detecting section 9 compares the item data of the keys of the records stored in the record registers 7 and 8 with the item 1 as a key based on the key information read from the record configuration information register 11. I do. As a result of the comparison by the break detection unit 9, if the item data of the key is the same, the break detection unit 9 does not perform any operation.
Then, the tallying calculation unit 10 performs the following tallying calculation using the tallying target data of the records stored in the second record register 8. That is, based on the information read from the tally calculation method instruction register 12 and the tally information read out from the record configuration information register 11, the tally calculation unit 10 calculates the tally target of records having the same key item data. An aggregation operation is performed using the data.

Also, as a result of the comparison by the break detecting section 9, when the key item data are different from each other, the break detecting section 9 detects a key break. For example, when the record 21d is input, the key item data of the records stored in the record registers 7 and 8 are different from each other (the key item data of the record 21d newly stored in the first record register 7). Is Kamakura, and the item data of the key of the record 21c newly stored in the second record register 8 is Fujisawa). For this reason, the break detection unit 9 detects a key break. Then, the break detection unit 9 sends the key break signal to the tally calculation unit 10.

When receiving the key break signal, the counting operation unit 10 counts the totaling data value 38 obtained by calculating the counting target data from the records 21a to 21c having the same key item data (Fujisawa), and the key item A record that associates the data (Fujisawa) with the data is output. In this manner, a plurality of records (record 21a to record 21) output from the data tabulation unit 2 shown in FIG.
f) is input to the data tabulation unit 3, each tabulation calculation result record (record 2) output from the data tabulation unit 3
2a and record 22b)
Is shown in FIG. At this time, based on the mask key information read from the record configuration information register 11, the counting operation unit 10 converts the item data of item 2 into the mask data AL based on the item data of the counting operation result record to be output.
L. Then, the tabulation calculation result records (record 22a and record 22b) are sent to the output control unit 5 and also to the data tabulation unit 4. Then, the counting operation unit 10 resets the counting operation, and calculates the counting target data of the record having the same item data as the item data of the key after the key break is detected (Kamakura, item data of item 2 does not matter). To perform a tally operation.

Then, the input record (record 2)
1a to record 21f), as described above,
Each operation by each unit is repeated.

(2-3-3) Operation of Data Tabulation Unit 4 Each record (record 22a, record 22b) output by tabulation calculation unit 10 of data tabulation unit 3 is stored in first record register 7 of data tabulation unit 4. Is entered. The functions of the record registers 7 and 8 in the data tabulation unit 4 are the same as the functions of the record registers in the data tabulation units 2 and 3.

The break detecting section 9 of the data totaling section 4 does not perform key break detection. For this reason, the data tabulation unit 4
The total calculation unit 10 performs total calculation using total data of all input records. Then, when all the records are input, the tally calculation unit 10 outputs a record in which the tally data calculated and the key item data are associated with each other. In this manner, when a plurality of records output from the data tabulation unit 3 shown in FIG. 3B are input, the tabulation calculation result record 23 output from the data tabulation unit 4 is shown in FIG. . At this time, the tallying operation unit 10 outputs the record 23 based on the mask key information read from the record configuration information register 11.
The item data of item 1 and item 2 are both mask data ALL.

In the operation of each of the data tabulation units 2, 3, and 4, the operation of the next data tabulation unit 3 starts after the operation of one data tabulation unit 2 ends, and the operation of the data tabulation unit 3 starts. The operation of the data totaling unit 4 does not start after the end. That is, the operations of the data tabulation units are performed simultaneously in parallel.

In this way, each data tabulation unit 2, 3, 4
Output each record (records 21a to 23).

(2-3-4) Operation of Output Control Unit 5 The output control unit 5 ranks the records (records 21 a to 23) output from the tally operation unit 10 of each data tally unit 2, 3, 4. While sorting in the order of the items, the totaling operation result database 97 as shown in FIG.

The operation of the output control unit 5 will be described in detail below. Tabulation operation unit 10 of each data tabulation unit 2, 3, 4
Are output in the following order. That is, the order of the records 21a, 21b, 21c, 21d, 22a, 21e, 21f, 22b, and 23 is as follows.
At this time, the order of the item data of item 1 of each record is Fujisawa, Fujisawa, Fujisawa, Kamakura, Fujisawa, Kamakura, Kamakura, Kamakura, and ALL. These records are sent from each of the data tabulation units 2, 3, and 4 to the output control unit 5. At this time, the output control unit 5 does not put together the records in the order output from each data summing unit, but puts them together as follows. That is, the output control unit 5 outputs the data
For each record output from 3 and 4, among the item data of each item, the mask data ALL is sorted so as to be last. Since the order of the item data of item 1 is reversed between the record 1d and the record 2a, the output control unit 5 determines that the record 1d and the record 2a
Swap the order between. Thus, the sorting process for item 1 ends. Looking at the item data of item 2 of each record, the records are sorted in the order of television, video, and stereo, so that the output control unit 5
No sort process is performed on item 2. As a result, FIG.
The output control unit 5 outputs a total calculation result database 97 as shown in FIG. Note that the output control unit 5
The total calculation result database 97 can be output as a whole, or can be output from the record 97a for each record of the total calculation result database 97.

In the present embodiment, each of the data tabulation units 2, 3, and 4 includes a key break detection process by the key break detection unit 6 and a tabulation calculation process by the tabulation calculation unit 10.
Thus, total data is generated for each item data of the key item. At this time, since the number of items of the key is different from each other, each data tabulation unit performs a different one-dimensional tabulation calculation process. The key break detection process and the tabulation calculation process performed by each data tabulation unit are relatively simple one-dimensional tabulation calculation processes. Therefore, the data tabulation units 2, 3, and 4 are configured with a simple hardware configuration. can do.

The plurality of data tabulation units 2, 3, and 4 constituting the data tabulation device 1 are connected in series in a descending order of the number of key items by making the number of items of each key different from each other. ing. Therefore, each of the data tabulation units 2, 3, and 4
A predetermined number of items (two items (item 1 and item 2), one item (item 1), and zero items (to calculate the totaled data of all input records)) specified in the order of the higher-order items Total data is generated for each item data. As a result, the data tally unit 2, which has a simple hardware configuration,
By simply connecting 3, 4 in series, it is possible to perform a multidimensional (in this case, three-dimensional) totalizing operation, which is a complicated totaling operation.

Further, since the data tabulation units 2, 3, and 4 are connected in series in descending order of the number of key items, the number of input records is reduced in the data tabulation unit 4 connected at the subsequent stage. . For this reason, the data tallying unit 4 connected at the subsequent stage can be configured to reduce its processing capability. As a result, the cost of the data totaling device 1 can be reduced. This cost reduction effect increases as the number of data tabulation units constituting the data tabulation device 1 increases.

In the present embodiment, however, it is not possible to cause each data tally unit to perform a tally calculation method (average value calculation or the like) that depends on the number of records to be tallyed. That is, since the data tabulation unit 3 and the data tabulation unit 4 only receive records of the result of the tabulation operation performed by the preceding data tabulation unit, the data tabulation unit 3 and the data tabulation unit 4 perform the tabulation calculation depending on the number of records. Can not do. In order for the data tallying device 1 to perform the tallying operation depending on the number of records, it is necessary to operate only the data tallying unit 2 without operating the data tallying unit 3 and the data tallying device 4. .

If the number of dimensions of the multidimensional totaling operation performed by the data totaling device 1 is smaller than the number of data totaling units in the data totaling device 1, only the data totalizing unit for the above-mentioned number of dimensions is operated. What should I do?

Embodiment 2 The data aggregation device according to the second embodiment performs a multidimensional aggregation operation using a plurality of records including n item data and corresponding aggregation target data. At this time, the plurality of records are sorted in advance by the sorting device in the order of higher-ranked items.

Also in the present embodiment, as an example of a multidimensional totaling operation, two item data (item data of item 1 and item data of item 2) and corresponding data to be totaled (item 3 The case where the three-dimensional totaling calculation is performed using a plurality of records consisting of

(3) Configuration of Data Aggregation Apparatus FIG. 4 is a block diagram showing a data aggregation apparatus according to the second embodiment. In the data totaling device shown in FIG. 4, the same components as those shown in FIGS. 1 and 2 are denoted by the same reference numerals, and description thereof will be omitted.

Data aggregation device 30 in the present embodiment
In this example, a plurality of data tabulation units 2, 3, and 4 are connected in parallel with different numbers of key items.

The output control unit 5 includes the data totaling units 2, 3,
The records output from the tallying operation unit 10 are sorted in the output order while being sorted in the order of the higher order items.

In the present embodiment, the number of data tabulation units arranged in the data tabulation device is not limited to three, and any number of data tabulation units can be arranged. For example, when performing a multidimensional totaling operation using a plurality of records including n items of data and corresponding target data, an (n + 1) -dimensional totaling operation is performed. In this case, (n + 1) data tabulation units are connected in parallel in the data tabulation device with the number of key items being different from each other.

The operation of the data collection device will be described below.

(4) Operation of Data Aggregation Apparatus The operation of the data aggregation apparatus configured as described above is as follows. That is, (4-1) an operation of storing necessary information in a record configuration information register and a tally calculation content register in response to an instruction from the outside;
(4-2) Operation for setting each unit based on the information stored in each register, (4-3) Sorted records are continuously input to data tabulation units 2, 3, and 4, and each data tabulation The operation of outputting the aggregation result records from the units 2, 3, and 4 is performed.

Note that (4-1) storage operation and (4-2)
Since the setting operation is the same as that of the first embodiment, the description thereof is omitted, and the operation of (4-3) will be described below. The detailed description of the same operation as that described in the first embodiment is omitted.

(4-3) Each record is successively input to each data tabulation unit 2, 3, 4 in the sorted order, and each data tabulation unit 2, 3, 4 outputs each tabulation result record. Operation The operation when the sorted records shown in FIG. 12 are input to the data tabulation units 2, 3, and 4 will be described.

(4-3-1) Operation of Data Tabulation Unit 2 The data tabulation unit 2 according to the present embodiment performs the same operation as the data tabulation unit 2 shown in the first embodiment. The operation different from the data tabulation unit 2 of the first embodiment is that the tabulation calculation result record output by the tabulation calculation unit 10 is output only to the output control unit 5.

(4-3-2) Operation of Data Tabulation Unit 3 The data tabulation unit 3 according to the present embodiment receives the sorted records in the same manner as the data tabulation unit 2 described above.
The key break detection process and the tallying calculation process performed on each input record are the same as those of the data tallying unit 3 of the first embodiment. Then, the counting operation unit 10
Is output to the output control unit 5 only.

(4-3-3) Operation of Data Tabulation Unit 4 The data tabulation unit 4 according to the present embodiment receives the sorted records in the same manner as the data tabulation unit 3 described above.
The tallying operation performed on each input record is the same as that of the data tallying unit 4 of the first embodiment.

(4-3-4) Operation of Output Control Unit 5 The output control unit 5 sorts the records output from the tally calculation units 10 of the data tally units 2, 3, and 4 while sorting the records in the order of higher-ranking items. A totaling operation result database as shown in FIG.

In the present embodiment, the sorted records are input to the data tabulation units 2, 3, and 4 in parallel. For this reason, the data tabulation units 2, 3, and 4 can independently perform one-dimensional tabulation calculation processing other than the sum value calculation calculation.

For example, the maximum value calculation operation, the minimum value calculation operation, and the average value calculation operation are instructed to the total calculation method instruction registers 12 of the data totaling units 2, 3, and 4 as information indicating different total calculation methods. The information is stored respectively.

Then, information for designating item 1 and item 2 as key target items is stored in the record configuration information register 11 of each of the data tabulation units 2, 3, and 4. Then, based on these pieces of information, the break detection unit 9 and the total calculation unit 10 of each of the data totaling units 2, 3, and 4 are set.

Thus, when the sorted records are continuously input to the data tabulation units 2, 3, and 4, the data output from the data tabulation units 2, 3, and 4 are the data of each record. The maximum value data of the aggregation target data corresponding to the key item data, the minimum value data of the aggregation target data corresponding to the key item data of each record, and the aggregation target data corresponding to the key item data of each record Is output. However, in this case, only three independent one-dimensional total calculation processes are performed, and no three-dimensional total calculation process is performed.

In this embodiment, as in the first embodiment, complicated aggregation calculation processing can be performed only by connecting data aggregation units having a simple hardware configuration in parallel.

Embodiment 3 In the case of performing a multi-dimensional totaling operation processing of four or more dimensions, if a plurality of (for example, four or more) data totaling units are arranged in one data totalizing apparatus, the scale of the apparatus becomes large. And the cost increases. Further, even if a plurality of data tallying devices in which three data tallying units are arranged are connected, similarly, the scale of the system increases and the cost increases.

The data totaling device according to the present embodiment is improved by improving the data totaling device according to the first embodiment.
Performs multi-dimensional aggregation calculation processing for the number of dimensions equal to or greater than the dimension. This makes it possible to perform multidimensional totaling calculation processing of four or more dimensions while reducing the scale of the data totaling device.

In the data totaling apparatus of the present embodiment, for example, a plurality of records having the following configuration are input. That is, item 1 (local name), item 2 (prefecture name), item 3
A plurality of records including (city name), item 4 (product name), and item 5 (data to be totaled) are input. In the present embodiment, a case will be described as an example of a multidimensional totaling operation in which a five-dimensional totaling operation is performed using the data to be totalized in the record. Each of the records is also sorted by the sorting device in the order of higher rank items before being input to the data totaling device.

(5) Configuration of Data Aggregation Device FIG. 5 is a diagram showing a configuration of a data aggregation device 40 according to the third embodiment. Data aggregation device 4 according to the third embodiment
0 is different from the data totaling apparatus 1 of the first embodiment in that an output unit 41 for outputting each totaling calculation result record is provided in the data totaling unit 4 connected at the last stage. The same components as those in FIG. 1 are denoted by the same reference numerals, and description thereof is omitted.

(6) Operation of Data Aggregation Device 40 The operation of the data aggregation device of the present embodiment will be described below. At this time, since the data totaling device 40 cannot perform the five-dimensional totaling operation at one time, first, the three-dimensional totalizing operation is performed, and the remaining two-dimensional totaling operation is performed using the totaling result.

(6A) First Three-Dimensional Tabulation Operation (6-1-1) First, each record is stored in the data tabulation device 40.
Before the data is input to the data aggregation unit 2, predetermined information is stored in the record configuration information register 11 and the aggregation calculation method instruction register 12 of each of the data aggregation units 2, 3, and 4.

That is, item 1, item 2, item 3, and item 4 are stored in the record configuration information register 11 of the data totaling unit 2.
Are key information indicating that the item 5 is a key, tabulation target information indicating that the item 5 is a tallying target, and mask key information indicating that there is no mask key target item.

The record configuration information register 11 of the data totaling unit 3 stores key information indicating that items 1, 2 and 3 are keys, totaling information indicating that item 5 is a totaling target, and item 4 Is stored as mask key information indicating that is a mask key target.

The record configuration information register 11 of the data totaling unit 4 stores key information indicating that items 1 and 2 are keys, totaling target information indicating that item 5 is a totaling target, items 3 and 4 Is mask key information indicating that the item is a mask key target item.

Each total calculation method instruction register 12 stores information instructing a total value calculation operation.

(6-1-2) Next, based on each information stored in each record configuration information register 11 and each totaling calculation method instruction register 12, each data totaling unit 2, 3, 4
Are set.

(6-1-3) Subsequently, each record is successively input to the data totaling unit 2 in the sort order. As a result, the data tabulation unit 2 generates a tabulation calculation result record for each item data of item 1, item 2, item 3, and item 4, and the data tabulation unit 3 generates the item 1, item 2, and item 3 A totaling operation result record is generated for each item data, and the data totaling unit 4 generates a totaling operation result record for each item data of item 1 and item 2. Then, each of the generated total calculation result records is output to the output control unit 5.
Output to At this time, the output control unit 5 holds the input totalization result records without outputting them. FIG. 6A shows a database 42 in which the aggregation calculation result records output from the output unit 41 are summarized.

(6B) Remaining Two-Dimensional Tabulation Operation (6-1-4) Subsequently, in order to perform the remaining two-dimensional tabulation calculation processing, the respective sections of the data tabulation units 2, 3, and 4 are set. cure. That is, the data tabulation unit 2 generates a tabulation calculation result record for each item data of item 1, and the data tabulation unit 3
The settings are reset so that a totaling calculation result record can be generated by performing a totaling calculation on the totaling target data of all the input records.

At this time, predetermined information is stored in the record configuration information register 11 and the total calculation method instruction register 12 of the data totaling units 2, 3, and 4, respectively. That is, the record configuration information register 11 of the data totaling unit 2 stores the item 1
Are key information indicating that item 5 is a key, counting target information indicating that item 5 is a counting target, and mask key information indicating that item 2 is a masking key. Thereby, the previously stored information is rewritten.

In the record configuration information register 11 of the data tabulation unit 3, key information indicating that there is no key target item, tabulation target information indicating that item 5 is a tabulation target, and items 1 and 2 are masked Mask key information indicating a key object is stored. Thereby, the previously stored information is rewritten.

The record configuration information register 11 and the tallying operation method instruction register 12 of the data tallying unit 4 store information that the data tallying unit is not operated. Thereby, the previously stored information is rewritten.

(6-1-5) Subsequently, based on each information stored in each record configuration information register 11 and each totaling calculation method instruction register 12, the data totaling units 2, 3, 4
Are set.

(6-1-6) Each tabulation result record generated for each item data of item 1 and item 2 output from the output section 41 of the data tabulation section 4 (FIG. 6 (a)
Are input to the data totaling unit 2, the data totaling unit 2 generates a totaling calculation result record for each item data of item 1, and the data totaling unit 3 generates all the input records. A totaling operation result record, which is a result of performing the totaling operation on the totaling target data, is generated, and the data totaling unit 4 does not generate the totaling operation result record. Then, each total calculation result record is output to the output control unit 5.

In this way, each totaling result record output from each data totaling unit 2, 3 is sent to the output control unit 5. The output control unit 5 sorts the totalized calculation result records and the totalized calculation result records held in (6-1-3) together in the order of the higher order items,
The totaling calculation result database 44 as shown in FIG.

As described above, in the data totaling apparatus of the present embodiment, by performing the totaling operation in a plurality of times, it is possible to perform the multidimensional totaling operation of an arbitrary number of dimensions.

Incidentally, there is also a method shown below for obtaining the database 44 of FIG. 6B. That is, the output control unit 5 outputs the totaling calculation result records output from each data totaling unit by the first three-dimensional totaling calculation operation to C
Output to PU or sorting device. Then, the output control unit 5 outputs each tally operation result record output from each data tally unit by the remaining two-dimensional tally operation to the CPU or the sorting device as it is. And the CPU
Alternatively, the sorting device may control the received total calculation result records to sort and combine them in the order of the higher-order items.

In this embodiment, the data totaling device 4
The output section 41 is provided in the data tabulation section at the last stage of “0”. For this reason, when performing multi-dimensional totaling calculation processing of four or more dimensions, due to limitations on cost and mounting area, etc.
Even when a plurality of data tabulation devices cannot be used, a multidimensional tabulation operation can be performed as follows. That is, by inputting each record output from the output unit 41 to the data totaling unit 2 of the data totaling device 40 again, it is possible to perform a multidimensional totaling calculation process of four or more dimensions.

Embodiment 4 The data tallying device according to the present embodiment is applied to a case where multidimensional tallying calculation processing of four or more dimensions is performed.

The data totaling device 50 according to the present embodiment includes:
A plurality of records having the following configurations are input. That is, item 1 (local name), item 2 (prefecture name), item 3 (city name), item 4 (product name), and item 5 (data to be tabulated)
Are input. In the present embodiment, as an example of a multidimensional totaling operation, a case will be described in which a five-dimensional totaling operation is performed using the data to be totalized of these records.

(7) Configuration of Data Aggregation Device FIG. 7 is a diagram showing the configuration of a data aggregation device 50 according to the fourth embodiment. Data totaling device 5 according to the fourth embodiment
0, the data totaling device 40 of the third embodiment
a and 40b are connected in series. The description of the same components as those in FIG. 5 is omitted.

In the present embodiment, there are two output control units 5a and 5b that receive each calculation result record.
These output control units 5a and 5b operate independently. Therefore, in the present embodiment, an output control device that receives both records output from the two output control units 5a and 5b is arranged. This output control device can combine the records output from the output control units while sorting the records in the order of the higher-ranking items.

(8) Operation of Data Aggregation Device The operation of the data aggregation device 50 according to the present embodiment will be described below.

(8-1-1) First, before each record is input, predetermined information is stored in the record configuration information register and the totaling operation content instruction register of each data totaling unit. That is, in the record configuration information register of the data totaling unit 2a, key information indicating that item 1, item 2, item 3, and item 4 are keys, aggregation target information indicating that item 5 is an aggregation target, and a mask Mask key information indicating that there is no key target item is stored.

The record configuration information register of the data totaling unit 3a stores key information indicating that items 1, 2, and 3 are keys, totaling information indicating that item 5 is a totaling target, and item 4 Mask key information indicating a mask key is stored.

In the record configuration information register of the data totaling unit 4a, key information indicating that items 1 and 2 are keys, totaling target information indicating that item 5 is a totaling target, and items 3 and 4 are stored. Mask key information indicating a mask key target item is stored.

In the record configuration information register of the data tabulation unit 2b, key information indicating that item 1 is a key, tabulation target information indicating that item 5 is a tabulation target, and item 2, item 3, and item 4 are stored. Mask key information indicating a mask key target item is stored.

In the record configuration information register of the data totaling unit 3b, key information indicating that there is no key target item, totaling target information indicating that item 5 is a totaling target, item 1, item 2, item 3 And mask key information indicating that item 4 is a mask key object.

The information for instructing the sum calculation is stored in the summation calculation instruction register of each data summation unit.

The record configuration information register 11 and the tallying operation method instruction register 12 of the data tallying unit 4b store information indicating that the data tallying unit 4b is not operated.

(8-1-2) Next, based on each information stored in each record configuration information register and each total calculation method instruction register 12, the break detection unit and total calculation unit of each data total unit are set. Is done.

(8-1-3) Then, when the records are successively input to the data tabulation unit 2a in the sort order, the data tabulation unit 2a outputs items 1, 2, 3, and 4
A totaling operation result record is generated for each item data of item 1, and the data totaling unit 3a generates item 1, item 2, and item 3
A totaling operation result record is generated for each item data item, the data totaling unit 4a generates a totaling operation result record for each item data of item 1 and item 2, and the data totaling unit 2b generates a totaling operation result record for item 1. Generation of a total calculation result record for each item data is performed. The data tabulation unit 3b generates a tabulation calculation result record by performing a tabulation operation using the tabulation target data of all input records. Note that the data tabulation unit 4b does not generate a tabulation calculation result record. Then, the tallying operation result records output from the data tallying units 2a, 3a, 4a are input to the output control unit 5a, and the data tallying unit 2b,
Each tabulation result record output from 3b is input to the output control unit 5b.

The output control units 5a and 5b sort the records in the order of output while sorting the records in the order of the higher order items. Then, the output control unit (not shown) receives the records output from each of the output control units 5a and 5b, and the output control unit outputs a tabulation calculation result database as shown in FIG. 6B.

As an alternative to the output control device,
A CPU or a sorting device may be used.

In this embodiment, a multidimensional totaling operation can be performed only by connecting a plurality of data totaling devices 40 in series. Also, as in Embodiment 4,
It is not necessary to change the setting of the aggregation operation once set for each data aggregation unit during the multidimensional aggregation operation.

Embodiment 5 FIG. The data totaling apparatus according to the fifth embodiment performs a multidimensional totaling calculation by a plurality of CPUs using a plurality of records including n item data and corresponding target data. That is, in the fifth embodiment, a multidimensional totaling operation is performed using a multiprocessor system. At this time, the plurality of records are sorted in advance by the sorting device in the order of higher-ranked items.

In the present embodiment, as an example of the multidimensional totaling calculation processing, a plurality of item data (item data of item 1 and item data of item 2) and a plurality of data to be totaled corresponding to these items are calculated. A case in which a three-dimensional aggregation operation is performed using records will be described.

(9) Configuration of Data Aggregation Device (Overall Configuration) FIG. 8 is a diagram showing the configuration of the data aggregation device according to the present embodiment. The data totaling device 60 according to the fifth embodiment includes a storage unit 61 storing a plurality (three) of data totaling instructions for instructing a plurality of (three) one-dimensional totaling operations, respectively, based on each data totaling instruction. CPUs 62, 63, and 64 that perform each one-dimensional aggregation operation, and each CPU
And an output control unit 65 that receives a totaling operation result record output from the server, performs predetermined control, and outputs the result. Each of the CPUs 62, 63, and 64 has a tally calculation counter that performs a tally calculation using tally data.

(Storage Unit 61) In the storage unit 61, when a plurality of items specified in the higher order are used as keys for two consecutively input records, the item data of the keys differ from each other. A key break detection step of detecting a key break, and a tabulation operation is performed on tabulation target data of records having the same key item data to generate tabulation data, and when a key break is detected, the tabulation data and the A totaling operation step of outputting a record in which the item data of the corresponding key is associated with (n +
1) (three in this embodiment) data totaling instructions are stored. Then, the number of key items specified in each data totaling instruction is different from each other (each CPU
62, 63, 64).

Each of the CPUs 62, 63, 64 decodes and executes each data totaling command. And each CPU
Reference numerals 62, 63, and 64 are connected in series in descending order of the number of key items of the data totaling command to be decoded and executed. In the present embodiment, a three-dimensional totaling operation is performed using a plurality of records including two item data (item data of item 1 and item data of item 2) and corresponding data to be totaled. Therefore, if three CPUs are connected in series, it is possible to perform a three-dimensional totaling calculation process.

In the present embodiment, the number of CPUs arranged in the data totaling device is not limited to three.
It is possible to arrange an arbitrary number of CPUs. For example, when performing a multidimensional totaling operation using a plurality of records including n items of data and corresponding target data, an (n + 1) -dimensional totaling operation is performed.
In this case, (n + 1) CPs
U are connected in series in ascending order of the number of key items, with the number of key items specified by each data totaling instruction to be decrypted and executed by each CPU differing from each other. Then, the storage unit 61 stores (n + 1) data totaling instructions.

(Output Control Unit 65) The output control unit 65 sorts the total calculation result records output from each of the CPUs 62, 63 and 64 in the output order while sorting them in the order of the higher-ranking items.

The operation of the data totaling apparatus 60 thus configured will be described below.

(10) Operation of Data Aggregation Device 60 (10-1) Before each record is input to the CPU 62, each of the CPUs 62, 63, and 64 first sends each data aggregation instruction from the storage unit 61. Read and decrypt. At this time, each of the CPUs 62, 63, and 64 is associated with each data totaling instruction (hereinafter, referred to as a first data totaling instruction, a second data totaling instruction, and a third data totaling instruction). Then, the number of key items instructed by the first data aggregation instruction is the first, the number of key items instructed by the second data aggregation instruction is the second largest, and the key instructed by the third data aggregation instruction Has the least number of items.

(Algorithm of Data Aggregation Instruction) The algorithm of each data aggregation instruction will be described with reference to the flowchart shown in FIG. In addition, the first data totaling instruction, the second
The data summarizing command and the third data summarizing command have the same algorithm, and differ only in the number of specified key items.

First, in step S102 (initialization step), each CPU performs an initialization process. In this initialization processing, the setting operation described in the first embodiment is performed.
At this time, each piece of information (key information, mask key information, and the like) is different from each other and stored in the storage unit 61. And
Each CPU reads information assigned to itself from the storage unit and performs a setting operation based on the information. At this time, each CPU resets each tally calculation counter.

In step S104 (each record input step), each CPU fetches each record that is continuously input.

In step S106 (key break detecting step), each CPU compares the item data of the keys of two consecutively input records with the item designated by the key information as a key. Then, as a result of the comparison, when the key item data are different from each other, a key break is detected, and the process proceeds to step S108. on the other hand,
As a result of the comparison, if the key item data is the same, the process proceeds to step S112.

In step S108 (total calculation step), each CPU outputs a total calculation result record that associates the total data calculated and the key item data. Then, control goes to a step S110.

At step S110 (total calculation step), each CPU resets the total calculation operation by each total calculation counter. Then, control goes to a step S112.

In step S112 (totaling operation step), each CPU causes the totalizing operation counter to perform a totalizing operation by using the totaling target data of the records having the same key item data.

In step S114, each CPU receives information indicating whether or not the input record still exists. If the input data still exists, each CPU proceeds to the operation in step S104. On the other hand, when there is no input data, each CPU ends the tallying calculation process.

(10-2) Each CPU performs each operation based on the decoded data totaling command. Each operation will be described below.

Before a record is input, each CPU
Each performs an initialization process. This initialization process is performed for each CP.
After the processing is performed by U, the tally calculation processing is performed by each CPU.

(10-2-1) Operation of CPU 62 When records are continuously input, the CPU 62 uses the item 1 and item 2 as keys, and sets the items of the mutual keys of the two continuously input records. Compare the data. As a result of the comparison, if they are the same, the CPU causes the tallying calculation counter to perform tallying calculation using the tallying target data of the input record.

As a result of the comparison, if they are different from each other, the CPU
Detect key breaks. Based on the detected key break, the CPU outputs a record in which the tabulated data, which is the result of the tabulation calculation by the tabulation calculation counter, and the key item data are associated with each other. This record is sent to the output controller 65 and also sent to the CPU 63. Then, the CPU 62 causes the tallying operation counter to reset the tallying operation, and uses the tallying operation counter of the record having the same item data as the item data of the key after the key break is detected. To perform an aggregation operation.

[0160] Each operation is performed by each unit as described above for continuously input records.

(10-2-2) Operation of CPU 63 When a record output from CPU 62 is input, C
The PU 63 compares the item data of the keys of the two input records with the item 1 as a key. As a result of the comparison, if they are the same, the CPU 63 causes the tallying calculation counter to perform tallying calculation using the tallying target data of the input record.

As a result of the comparison, if they are different from each other,
Detects a key break. Based on the detected key break, the CPU 63 outputs a record in which the tabulated data, which is the result of tabulated calculation by the tabulation calculation counter, and the key item data are associated with each other. At this time, CP
U63 sets ALL of the item data of item 2 as the item data of the record to be output. This record is sent to the output control unit 65 and also to the CPU 64. Then, the CPU 63 causes the tallying operation counter to reset the tallying operation, and uses the tallying target data of the record having the same item data as the item data of the key after the key break is detected, to the tallying operation counter. Causes a new aggregation operation to be performed.

Then, as described above, each operation is performed on each of the continuously input records.

(10-2-3) Operation of CPU 64 The record output from the CPU 63 is input.
In PU64, since the items to be keyed are not set,
No key break detection. For this reason, the CPU 64
In the calculation, a totaling operation is performed using totaling target data of all the input records. When all the records have been input, the CPU 64 outputs a record in which the tabulated data, which is the result of the tabulation operation, is associated with the key item data. At this time, the CPU 64 determines that the item data of item 1 is AL
L, the item data of item 2 is ALL.

In this way, each record is output from each data totaling unit.

In the operation of each of the CPUs 62, 63, and 64, the operation of the next CPU 63 starts after the operation of one CPU data 62 ends, and the operation of the CPU 64 starts after the operation of the CPU 63 ends. is not. That is, the operations of the CPUs 62, 63, and 64 are simultaneously performed in parallel.

(10-2-4) Operation of Output Control Unit 65 The output control unit 65 sorts the totalized calculation result records output from each of the CPUs 62, 63 and 64 in the order of the higher order items, and collects them in the output order. , And outputs a total calculation result database as shown in FIG.

In the present embodiment, the respective data totaling instructions that are specified so that the number of items of each key are different from each other are executed by the respective CPUs 62, 63, and 64, respectively.
For this reason, the CPUs 62, 63, and 64 execute the respective data totaling instructions, instead of executing the data totaling instruction sequence in which all the data totaling instructions are collected by one CPU as in the related art. As a result, the load on each CPU is reduced, and the processing speed of each CPU is increased.

Further, since the CPUs are connected in series in descending order of the number of keys of the key of the data totaling command to be decoded and executed, the number of input records becomes smaller as the CPU is connected at a later stage. For this reason, C
It can be configured to reduce the processing power of the PU. In addition, the processing capacity of each CPU may be set to be the same, and the subsequent CPU may perform processing other than the tallying calculation processing. As a result, the load on each CPU can be made uniform.

Embodiment 6 FIG. FIG. 10 is a diagram illustrating a configuration of a data totaling device according to the present embodiment. FIG.
The same components as those described above are denoted by the same reference numerals, and detailed description thereof will be omitted. The data aggregation device 70 according to the sixth embodiment differs from the data aggregation device 60 according to the fifth embodiment in that the CPUs 62, 63, and 64 are connected in parallel. The output control unit 65 includes the CPUs 62 and 6
The totaling calculation result records output from 3, 64 are put together in the output order while being sorted in the order of the higher order items.

(Operation of Data Aggregation Apparatus) Each CPU 62
63 and 64 are each data totaling instruction (first
(Data totaling instruction, second data totaling instruction, and third data totaling instruction), and decodes and executes each data totaling instruction.

First, before records are successively input in the sort order, each CPU performs an initialization process. After this initialization process is performed by each CPU, each C
Aggregation calculation processing is performed by the PU.

The operation when each of the sorted records shown in FIG. 12 is input to each CPU will be described. In addition,
The description of the same operations as those described in the fifth embodiment is omitted.

(11-1) Operation of CPU 62 The same operation as that of the CPU 62 described in the fifth embodiment is performed in the CPU 62 according to the present embodiment. The operation different from that of the CPU 62 of the fifth embodiment is that the total calculation result record output by the CPU 62 is output only to the output control unit 65.

(11-2) Operation of CPU 63 As in the case of the CPU 62, the sorted records are input to the CPU 63 of the present embodiment. The key break detection process and the tallying calculation process performed for each input record are the same as those of the fifth embodiment.
It is the same as U63. Then, the aggregation calculation result record output by the CPU 63 is output only to the output control unit 65.

(11-3) Operation of CPU 64 The data CPU 64 of the present embodiment includes the CPU
As in the case of 62 and 63, the sorted records are input. The tabulation calculation process performed on each input record is the same as that of the CPU 64 of the fifth embodiment.

(11-4) Operation of Output Control Unit 65 The output control unit 65 sorts the records output from the CPUs 62, 63 and 64 in the order of the higher-ranking items and collects them in the order of the output. And outputs a total calculation result database as shown in FIG.

Also in this embodiment, each CPU 62,
Since 63 and 64 execute each data totaling instruction, each CP
The load on U is reduced, and the processing speed of the CPU is increased.

[0179]

According to the first aspect of the present invention, the plurality of data tabulation means constituting the data tabulation device make the number of items of each key different from each other, so that the number of key items is in descending order. , Are connected in series. For this reason, first, in the data totaling means in which each record is input in the sort order, the total data is generated for each item data of the n items specified in the order of the higher order items. Next, the data tabulation means outputs a record in which the tabulated data is associated with the item data of the key corresponding to the data to the data tabulation unit at a subsequent stage connected to the data tabulation means. Subsequently, the data tabulation means generates tabulated data for each item data of the (n-1) items specified in the order of the higher order items. Then, the data tabulation means outputs a record in which the tabulated data is associated with the item data of the key corresponding to the tabulated data to the data tabulation means at a subsequent stage connected to the data tabulation means. In this way, the plurality of data tallying units constituting the data tallying device respectively generate tally data for each item data of the plurality of items.

[0180] For this reason, the tallying calculation process performed by the entire data tallying device is complicated.

On the other hand, the processing performed by each data totaling means only generates total data for each item data of the key item. The key break detection process and the tallying calculation process performed at the time of this generation are relatively simple processes.
For this reason, each data aggregation means can be configured with a simple hardware configuration.

Therefore, a complicated tallying operation can be performed only by connecting data tally means having a simple hardware configuration in series.

Further, since the data totaling means are connected in series in the descending order of the number of key items, the number of records to be input becomes smaller as the data totaling means is connected at a later stage. For this reason, it is possible to configure so as to reduce the processing capacity of the data aggregation means connected at the subsequent stage. As a result, the cost of the data totaling device can be reduced.

According to the second aspect of the present invention, the plurality of data tabulation means constituting the data tabulation device are connected in parallel with each key having a different number of items.
Then, the plurality of data totaling means generates total data for each item data of the plurality of items specified in the order of the higher order items. As a result, complicated aggregation calculation processing can be performed by simply connecting data aggregation means having a simple hardware configuration in parallel.

According to the third aspect of the present invention, each record output from each totaling operation means of the data totaling device is controlled by the output control means. Therefore, regardless of the order in which the records output from the tallying calculation means are output, the output control means can sort the records while sorting them in the order of higher-ranking items.

According to the invention described in claim 4 or 5, each data totaling command having a different number of key items is:
Each is executed by each CPU. For this reason, one CPU does not execute a data totaling instruction sequence in which all data totaling instructions are integrated as in the prior art, but each CPU executes each data totaling instruction. It is reduced and the processing speed of the CPU is increased.

According to the sixth aspect of the present invention, each CPU
Each record output from is controlled by output control means. Therefore, regardless of the order in which the records output from the CPUs are output in any order, the output control means can sort the records while sorting them in the order of the higher-ranked items.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a configuration of a data aggregation device according to a first embodiment;

FIG. 2 is a diagram showing a configuration of a data tabulation unit according to the first embodiment.

FIG. 3A is a diagram illustrating a database output from a data totaling unit 2 according to the first embodiment, and FIG.
FIG. 3C is a diagram illustrating a database output from the data totaling unit 3 according to the first embodiment, and FIG.
FIG. 6 is a diagram showing a database output from a data totaling unit 4 which is a database.

FIG. 4 is a diagram illustrating a configuration of a data aggregation device according to a second embodiment;

FIG. 5 is a diagram showing a configuration of a data aggregation device according to a third embodiment.

FIG. 6A is a diagram illustrating a database output from an output unit according to the third embodiment, and FIG. 6B is a diagram illustrating a database output from an output control unit;

FIG. 7 is a diagram illustrating a configuration of a data aggregation device according to a fourth embodiment.

FIG. 8 is a diagram illustrating a configuration of a data aggregation device according to a fifth embodiment.

FIG. 9 is a flowchart illustrating an algorithm of a data totaling instruction stored in a storage unit according to the fifth embodiment.

FIG. 10 is a diagram showing a configuration of a data aggregation device according to a sixth embodiment.

FIG. 11 is a diagram showing a configuration of a conventional system for performing a tallying operation.

FIG. 12 is a diagram showing a database sorted by a conventional sorting device.

FIG. 13A is a diagram showing a total data value stored in each conventional total operation counter. (B)
FIG. 4 is a diagram showing a database output from a CPU.

[Explanation of symbols]

1, 30, 40, 40a, 41a, 50, 60, 70
Data aggregation device, 2, 2a, 2b, 3, 3a, 3b,
4,4a, 4b Data tabulation unit, 5,5a, 5b, 65
Output control unit, 6 key break detection unit, 7 first record register, 8 second record register, 9 break detection unit, 10 total operation unit, 11 record configuration information register, 12 total operation method instruction register, 21, 4
2,44,92,97 database, 21a-21
f, 22a, 22b, 23, 92a to 92q, 97a to
97f records, 61 storage units, 62, 63, 64,
82 CPU, 80 sorting device, 93, 94, 95 totaling operation counter.

Claims (6)

[Claims] 1. A plurality of records, each of which includes n (n is an integer) item data and corresponding data to be totaled, are sorted in the order of the higher-ranking items, and are totalized using the totalized data of each sorted record. In a data aggregation device that performs an operation, a key that detects a key break in which item data of different keys are different when two or more items specified in a higher order are used as keys for two consecutively input records A break detecting means for calculating tabulated data of records having the same key item data to generate tabulated data, and, when a key break is detected, the tabulated data and the corresponding key item data And (n + 1) data totaling means having a function of outputting a record in which the number of items of each key are different from each other. , The data totaling device being connected in series in descending order of the number of key items. 2. A plurality of records each including n (n is an integer) item data and corresponding data to be totaled are sorted in the order of the higher order items, and are totalized using the totalized data of each sorted record. In a data tabulation device that performs an operation, when two or more records specified consecutively are used as keys for two records that are consecutively input in sort order, a key break in which the item data of each key is different is detected. Key break detection means, and tabulation calculation of data to be tabulated of records having the same key item data to generate tabulated data, and when a key break is detected, the tabulated data and the key corresponding thereto are detected. (N + 1) data tabulation means having a function of outputting a record in which item data is associated with each other. A data totaling device characterized by being differently connected in parallel. 3. The data according to claim 1, further comprising an output control unit that sorts the records output from the tallying calculation unit of each of the data tallying units and sorts them in the order of output while sorting them in the order of higher-ranking items. Aggregation device. 4. A plurality of records each including n (n is an integer) item data and corresponding data to be totaled are sorted in the order of the higher-ranked items, and are totalized using the totalized data of each sorted record. In a data aggregation device that performs an operation, a key that detects a key break in which item data of different keys are different when two or more items specified in a higher order are used as keys for two consecutively input records A break detection step, and a tabulation operation is performed on tabulation target data of records having the same key item data to generate tabulation data, and when a key break is detected, the tabulation data and the corresponding key item data are generated. A storage means for storing (n + 1) data totaling instructions having a totaling operation step of outputting a record corresponding to and (n + 1) CPUs for decoding and executing (n + 1) data totaling instructions, respectively, wherein the (n + 1) data totaling instruction strings have different numbers of key items from each other; The (n + 1) CPUs are serially connected in descending order of the number of key items of a data totaling command to be decoded and executed. 5. A plurality of records including n (n is an integer) item data and corresponding data to be totaled are sorted in the order of the higher order items, and are totaled using the totalized data of each sorted record. In a data tabulation device that performs an operation, when two or more records specified consecutively are used as keys for two records that are consecutively input in sort order, a key break in which the item data of each key is different is detected. A key break detection step to perform a tabulation operation on tabulation target data of records having the same key item data to generate tabulation data, and, when a key break is detected, the tabulation data and a key corresponding thereto are detected. A totaling operation step of outputting a record corresponding to the item data, and (n + 1) data totaling instructions stored therein. Means, and (n + 1) CPUs for decoding and executing (n + 1) data totaling instructions, respectively, wherein the (n + 1) data totaling instructions have different key item numbers from each other. The data aggregation device, wherein the (n + 1) CPUs are connected in parallel. 6. The data totaling device according to claim 4, further comprising an output control unit that sorts the records output from the CPUs and sorts the records in the order of higher rank items, and sorts the records in the order of output.