JP2017037511A - Analyzer, analyzing method, and analyzing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an analyzer capable of carrying out a series of analysis processing at a high speed by reducing the number of access to a table storing records of analyzing objects even when attributes of plural selection conditions are different from each other.SOLUTION: A multi query generation section 51 receives specifications of a plurality of selection conditions each having an attribute different from each other, and generates a query for each of the selection conditions. A query tree processing section 522 determines whether a record stored in the storage 4 matches with the respective selection conditions corresponding to the plural queries generated by the multi query generation section 51 and analyzes the record which matches with the selection condition on each selection condition.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an analysis apparatus, an analysis method, and an analysis program.

  In recent years, the importance of technology for extracting useful information from a large amount of data called big data has increased. For example, when the sales strategy is decided in consideration of the influence of the region and time of sales data, the analysis result expressed by the group data (group-by) and the aggregation function (aggregation) is used. In many cases, partial data that produces analysis results with a large discrepancy between the two is useful.

  Here, the partial data means a group of records extracted from the entire data by a query that performs selection processing according to the selection condition represented by A = 1 when the value of the attribute A is l. In addition, partial data that produces analysis results with a large discrepancy are defined as follows. That is, the result of executing the analysis instruction q on the table D as the entire data is represented as q (D), and the result of executing the analysis instruction q on the arbitrary partial data v (D) of the table D is represented by q *. v (D). In this case, the partial data that produces an analysis result with a large divergence means v (D) that maximizes the divergence between q (D) and q * v (D). Defined.

  Here, U (A, B) is a function for calculating the difference between A and B. U (A, B) is a function that can calculate a difference with respect to input data D. For example, the Euclidean distance can be distributed between D and Δd when the input data is D and the difference input data is Δd as shown in the following equation (2), that is, the exchange law and the coupling law are established. It is a function and can be used as a function for calculating the deviation.

  In order to specify the partial data that produces the analysis result with a large difference expressed as described above, for the analysis command q for the arbitrary partial data v (D), q (D) and q * v (D) It is necessary to calculate the degree of divergence, but this takes an enormous amount of time.

  Conventionally, when a query that performs a plurality of selection processes is executed, each query accesses the table D. Therefore, there is a problem that the execution time increases linearly as the number of queries to be executed increases. Therefore, a technique is known that reduces the number of accesses to the table by sharing the selection process when the attribute A of the selection conditions of the plurality of selection processes is the same. For example, by performing equivalent conversion of a plurality of selection processing queries having the same selection condition attribute A specified in the selection processing using the order by clause into one query, the database management system (DBMS) is used to Processing is performed (see Non-Patent Document 1).

Yohei Mizuno, Yoshihiro Kishida, Yuki Arase, Toshimori Honjo, Makoto Onizuka, “Efficient search of partial data that produces highly useful analysis results”, 2015, DEIM Forum 2015 D6-4

  However, when the attribute A of each selection condition specified in a plurality of selection processes is different, the selection process cannot be shared, and the number of accesses to the table cannot be reduced.

  The present invention has been made in view of the above, and reduces the number of accesses to a table in which records to be analyzed are stored even when attributes of a plurality of selection conditions are different, thereby speeding up analysis processing. For the purpose.

  In order to solve the above-described problems and achieve the object, the analysis apparatus according to the present invention accepts designation of a plurality of selection conditions having different attributes, and generates a query for each selection condition, and a storage unit An analysis unit that determines whether each selection condition corresponding to a plurality of queries generated by the generation unit is satisfied for each record to be stored, and analyzes a record that matches the selection condition for each selection condition And.

  According to the present invention, even when the attributes of a plurality of selection conditions are different, the number of accesses to a table storing records to be analyzed can be reduced to speed up the analysis process.

FIG. 1 is a schematic diagram showing a schematic configuration of an analyzer according to an embodiment of the present invention. FIG. 2 is a diagram illustrating a data configuration of the analysis target table according to the present embodiment. FIG. 3 is a schematic diagram illustrating the query tree of this embodiment. FIG. 4 is an explanatory diagram for explaining a multiple selection process according to the present embodiment. FIG. 5 is an explanatory diagram for explaining the function of the query tree processing unit of the present embodiment. FIG. 6 is an explanatory diagram for explaining the degree of deviation between the analysis result of partial data and the analysis result of overall data according to the present embodiment. FIG. 7 is a diagram illustrating an example of a sales amount histogram in a record group that has not been subjected to the multiple selection processing according to the present embodiment. FIG. 8 is a diagram illustrating a histogram of customer ages in a record group that has not been subjected to the multiple selection process of the present embodiment. FIG. 9 is a flowchart showing the multiple selection processing procedure of the present embodiment. FIG. 10 is a flowchart showing the deviation determination processing procedure of the present embodiment. FIG. 11 is a diagram illustrating a computer that executes an analysis program.

  Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. In addition, this invention is not limited by this embodiment. Moreover, in description of drawing, the same code | symbol is attached | subjected and shown to the same part.

[Analysis equipment]
FIG. 1 is a schematic diagram illustrating a schematic configuration of an analyzer according to the present embodiment. The analysis device 1 is realized by a general-purpose computer such as a workstation or a personal computer, and includes an input unit 2, an output unit 3, a storage unit 4, and a control unit 5.

  The input unit 2 is realized by using an input device such as a keyboard or a mouse, and inputs various instruction information such as processing start to the control unit 5 in response to an input operation by a data analyst. The output unit 3 is realized by a display device such as a liquid crystal display, a printing device such as a printer, an information communication device, and the like, and outputs a result of analysis processing described later to a data analyst.

  The storage unit 4 is realized by a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk, and stores a table D to be analyzed later. The storage unit 4 may be configured to communicate with the control unit 5 via an electric communication line such as a LAN or the Internet.

  FIG. 2 is a diagram illustrating a data configuration of the table D. As illustrated in FIG. As shown in FIG. 2, in the table D, each attribute value 1 of the plurality of attributes A is specified for each record. FIG. 2 shows an order table D, and the attribute A includes an order number, sales amount, order date, store name, product genre, customer age, and customer sex. For example, for the record of order number 1, the sales amount is 8000 yen, the order date is April 2015, the store name is Osaka, the product genre is clothing, the customer age is 45, and the customer gender is male. ing.

  As illustrated in FIG. 1, the control unit 5 performs a plurality of query generation units 51 and a plurality of query execution units 52 by executing a processing program stored in a memory by an arithmetic processing unit such as a CPU (Central Processing Unit). Function.

  The multiple query generation unit 51 receives designation of a plurality of selection conditions having different attributes, and generates a query for each of the selection conditions. Specifically, the multiple query generation unit 51 executes a selection process for acquiring partial data according to a selection condition for specifying an attribute, based on the attribute A of the table D that is the target of the analysis instruction input by the data analyst. Generate a query.

  For example, a plurality of queries are generated corresponding to a plurality of selection conditions exemplified in the following formula (3). In the present embodiment, a query for acquiring partial data that matches the selection condition 1 with the customer gender as male, a query for acquiring partial data that matches the selection condition 2 with the customer age 40s, and the product genre A query for acquiring partial data that matches the selection condition 3 with clothes as a clothing, a query for acquiring partial data that matches the selection condition 4 with goods genre as miscellaneous goods, and the like are generated.

  The multiple query execution unit 52 includes a query tree construction unit 521 and a query tree processing unit 522. The query tree construction unit 521 constructs a query tree having the attribute A and the attribute value l specified by each selection condition as nodes from the query generated by the multiple query generation unit 51. Specifically, the query tree construction unit 521 stores the plurality of queries generated by the multiple query generation unit 51 as a query tree and the specified selection condition in the cache memory as key information.

  FIG. 3 is a diagram illustrating a configuration of a query tree. As shown in FIG. 3, in the query tree, each selection condition is represented by a path connecting the attribute A and its attribute value l. FIG. 3 shows selection condition 1 for male sex, selection condition 2 for customer age 40, selection condition 3 for product genre, and selection condition 4 for product genre as miscellaneous goods. A query tree composed of four selection conditions is illustrated.

  The query tree processing unit 522 performs analysis processing including multiple selection processing and divergence determination processing. In the multiple selection process, the query tree processing unit 522 determines for each record of the table D stored in the storage unit 4 whether the selection conditions corresponding to the multiple queries generated by the multiple query generation unit 51 are met. For each selection condition, the analysis instruction q for the record that matches the selection condition is executed. As a result, an analysis result of partial data that is a set of records that match each selection condition is acquired.

  Specifically, in the multiple selection process, the query tree processing unit 522 refers to the table D that is the target of the analysis instruction q in the storage unit 4, and the attribute value l of the attribute A that matches the query tree selection condition for each record. Check if there is any. If they match, the query tree processing unit 522 executes the analysis instruction q for the record that has been subjected to the multiple selection processing having the same attribute value l, and stores the analysis result for the record in an appropriate memory area. . The query tree processing unit 522 performs the above processing for all the records in the table D.

  Here, referring to FIGS. 2 to 5, for example, when the analysis instruction q is expressed by the following equation (4), that is, for the order table D, the total sales amount is calculated for each order year and month. The case where it is is demonstrated.

  In this case, the order number 1 record in the order table D shown in FIG. 2 matches the selection conditions 1 to 3 of the query tree shown in FIG. Therefore, as shown in the following equation (5), the query tree processing unit 522 executes the analysis instruction q for the query tree selection conditions 1 to 3, and records 8000 yen in the sales amount in April 2015. The total sales amount is 8,000 yen. Then, as illustrated in FIG. 4, the query tree processing unit 522 stores the analysis result in association with each path of the query tree, that is, the corresponding selection conditions 1 to 3.

  Next, when the value 1 of the attribute A is confirmed for the record of the order number 2, it does not match any of the selection conditions 1 to 4 of the query tree shown in FIG. Therefore, the search unit 522 removes the record of the order number 2 from the search, and moves the process to the next record.

  Next, when the value 1 of the attribute A is confirmed for the record with the order number 3, only the selection condition 1 of the query tree shown in FIG. Therefore, as shown in the following equation (6), the query tree processing unit 522 executes the analysis instruction q only for the selection condition 1 of the query tree, and adds 4000 yen to the sales amount in April 2015 to calculate the sales amount. The total of 12,000 yen. Then, as illustrated in FIG. 5, the query tree processing unit 522 stores the analysis result in association with the path of the query tree, that is, the corresponding selection condition 1. For the query tree selection conditions 2 and 3, the total sales amount in April 2015 remains 8000 yen.

  As described above, according to the multiple selection processing, the query tree processing unit 522 performs analysis after executing a plurality of selection processing queries specified by different selection conditions of the plurality of attributes A at a time for each record of the table D. The instruction q is executed, and an analysis result of partial data that matches each selection condition is obtained. Thereby, the analysis process can be speeded up.

  In the divergence determination process, the query tree processing unit 522 performs analysis of partial data that matches each selection condition and the entire data before the above-described multiple selection process for each record is completed for all the records in the table D. The degree of divergence from the analysis result is calculated, and the multiple selection processing for the subsequent records is canceled for the selection condition where the degree of divergence cannot be maximized.

  Here, with reference to FIG. 6, the degree of deviation between the analysis result of the partial data and the analysis result of the entire data will be described. FIG. 6 shows an analysis result of the entire data of the table D as an analysis result of executing the analysis instruction q shown in the above formula (4) of the partial data corresponding to the selection conditions 1 to 3 shown in the above formula (3). The degree of deviation from the distance, that is, the Euclidean distance is shown. The x-axis in FIG. 6 indicates the order date and the y-axis indicates the normalized sales amount.

  Here, the Euclidean distance between the analysis result of the entire data and the analysis result of the partial data of the selection condition 1 is represented by the sum of the Euclidean distances for each month (see the above formula (2)). Further, the Euclidean distance for each month is represented by the difference in the normalized sales amount for each month.

  In FIG. 6, the difference between the normalized sales price of all products in each month and the normalized sales price of men, that is, the Euclidean distance U (all products, men) of each month is 0 in January 2015. Suppose that January and February are 0.04, March is 0.19, and April is 0.02. In this case, the Euclidean distance U (all products, male) between the analysis result of the entire data and the analysis result of the partial data of the selection condition 1 is Euclidean distance 0.1 in January 2015 and Euclidean distance 0 in February. .04, March Euclidean distance 0.19 and April Euclidean distance 0.02 are calculated as 0.35. Similarly, the Euclidean distance U (all products, 40s) between the analysis result of the entire data and the analysis result of the partial data of the selection condition 2 is calculated as 0.30. Further, the Euclidean distance U (all products, clothes) between the analysis result of the entire data and the analysis result of the partial data of the selection condition 3 is calculated as 0.74. Therefore, it can be understood that the partial data with the maximum deviation degree satisfying the following expression (7) is the partial data of the clothing of the selection condition 3.

  When calculating the degree of divergence, the query tree processing unit 522 of the present embodiment acquires the statistical information of all records and the multiple selection process for each record before completing all the records in the table D. The statistical information of the unprocessed record that matches each selection condition is derived using the statistical information of the processed record that matches each selection condition, and the degree of deviation is calculated using the statistical information of the unprocessed record. calculate.

  Specifically, prior to the divergence determination process, the query tree processing unit 522 has a histogram showing the distribution of the value 1 of the attribute A subject to the analysis instruction q for all record groups in the table D, that is, the entire data, and each part. A histogram indicating the distribution of the value l of the attribute A specified by the data selection condition is created. For example, a histogram showing the distribution of the attribute of the analysis command q shown in the above formula (4), that is, the value of the sales amount, is created. Further, a histogram showing the distribution of each attribute value 1 of the selection condition 1 attribute shown in the above formula (3), that is, the customer gender, the selection condition 2 attribute, ie, the customer age, and the selection condition 3, ie, the product genre Keep it.

  In the divergence determination process, the query tree processing unit 522 first describes the analysis result of partial data that matches each selection condition (hereinafter referred to as q * selection condition (D1)) for the record group D1 that has been subjected to the multiple selection process. ) And the analysis result q (D1) of D1 is calculated. In this case, as the degree of deviation, for example, it is assumed that the Euclidean distance is calculated as shown in the following equation (8).

  Here, the following equation (9) is established for the function for calculating the degree of deviation.

  Therefore, when the query tree processing unit 522 performs the multiple selection process on the unprocessed record group D2 using the histogram of the attribute values 1 of the record group D2 in which the multiple selection process for each record is not processed, The upper limit value and the lower limit value of the degree of deviation between the analysis result of the partial data that matches each selection condition and the analysis result of the entire data are calculated. For simplification of explanation, it is assumed that the target attribute value l of the analysis instruction q takes a positive value. However, when the target attribute value l of the analysis instruction q takes a negative value, it can be handled in the same manner by converting the value range to positive.

  First, the query tree processing unit 522 creates a histogram indicating the distribution of the value l of each attribute A of the record group D1 that has been subjected to the multiple selection processing for each record. Then, the histogram of each attribute value l of the unprocessed record group D2 is derived by subtracting the histogram of the processed record group D1 from the histogram of the entire data. For example, FIG. 7 is a histogram illustrating the distribution of the value of the attribute A that is the object of the analysis command q, that is, the sales amount, for the unprocessed record group D2. FIG. 8 is a histogram illustrating the distribution of customer age values for the unprocessed record group D2.

  The query tree processing unit 522 acquires the number of records having the value 1 of the attribute A of each selection condition from the histogram of the unprocessed record group D2. For example, according to FIG. 8, it can be seen that the number of unprocessed records for selection condition 2, that is, customer age 40s, is two.

  When the unprocessed record acquired here takes the maximum value, the analysis result of the unprocessed record group D2 and the analysis result of the partial data that matches each selection condition of the unprocessed record group D2 The degree of deviation is the upper limit. Therefore, the query tree processing unit 522 extracts the attribute value l of the number of unprocessed records from the histogram of the value l of the attribute A to be analyzed in the unprocessed record group D2.

  For example, according to FIG. 7, if two attribute values 1 are extracted in descending order of sales amount, 10000 yen is two. Thereby, the upper limit value of the degree of deviation between the analysis result of the unprocessed record group D2 and the analysis result of the partial data that matches the selection condition 2 in the unprocessed record group D2 is calculated as 10000 + 10000 = 20000. . Therefore, the upper limit value of the degree of divergence between the analysis result of Table D and the analysis result of the partial data that matches the selection condition 2 refers to the Euclidean distance of the above equation (8), as shown in the following equation (10): , 120,000+ (10000 + 10000) = 14,000.

  Similarly, when the unprocessed record having the value l of the attribute A of each selection condition takes the minimum value, the analysis result of the unprocessed record group D2 and each selection condition of the unprocessed record group D2 The degree of deviation from the analysis result of the partial data that matches is the lower limit. Therefore, the query tree processing unit 522 extracts the attribute values of the number of unprocessed records from the histogram of the value 1 of the attribute A to be analyzed in the unprocessed record group D2.

  For example, when the number of records having the value 1 of the attribute A of the selection condition 1 in the unprocessed record group D2 is 3, according to FIG. 7, three attribute values l are extracted in ascending order of sales amount. Then, 4000 yen is 2 cases and 6000 yen is 1 case. Therefore, the lower limit value of the degree of deviation between the analysis result of the unprocessed record group D2 and the analysis result of the partial data matching the selection condition 1 in the unprocessed record group D2 is calculated as 4000 + 4000 + 6000 = 14000. Therefore, the lower limit value of the degree of deviation between the analysis result of Table D and the analysis result of the partial data that matches the selection condition 1 refers to the Euclidean distance of the above equation (8), and is expressed by the following equation (11): , 250,000+ (4000 + 4000 + 6000) = 264000.

  The query tree processing unit 522 compares the lower limit value and the upper limit value of the degree of divergence between the analysis results of the partial data and the analysis results of the whole data that match a plurality of selection conditions, so that the degree of divergence is maximized. Identify the selection conditions that cannot be. For example, if the lower limit of the degree of divergence of partial data that meets a certain selection condition is greater than the upper limit of the degree of divergence of partial data that meets other selection conditions, the latter degree of divergence is It can't be. In the above example, the upper limit value of the degree of divergence of the partial data that matches the selection condition 2 is less than the lower limit value of the degree of divergence of the partial data that matches the selection condition 1, so the partial data that matches the selection condition 2 It can be seen that the degree of deviation cannot be the maximum value. In this case, the query tree processing unit 522 stops the process for the specified selection condition 2 in the multiple selection process for subsequent records.

  As described above, according to the divergence determination process, when the multiple selection process for each record is performed on the unprocessed record group D2 before the multiple selection process for each record is completed for all the records in the table D, It can be seen whether or not the degree of divergence between the analysis result of the partial data that matches each selection condition and the analysis result of the entire data can be maximized. Therefore, the query tree processing unit 522 cancels the subsequent multiple selection processing of records for a selection condition in which the degree of divergence between the analysis result of partial data that matches each selection condition and the analysis result of the entire data cannot be maximized. . Thereby, it is possible to speed up the analysis process and search for a selection condition for selecting partial data having a high degree of deviation from the analysis result of the entire data.

[Analysis processing]
Next, the analysis processing procedure in the analyzer 1 will be described with reference to the flowcharts of FIGS. 9 and 10. First, FIG. 9 is a flowchart showing a multiple selection processing procedure. The flowchart in FIG. 9 starts at the timing when an analysis start instruction is input via the input unit 2 by the data analyst, for example.

  First, the query tree processing unit 522 reads the query tree stored in the cache memory (step S1), refers to the table D that is the target of the analysis instruction q in the storage unit 4, and selects a record that has not undergone multiple selection processing. The presence or absence is confirmed (step S2), and if there is an unprocessed record (step S2, Yes), the unprocessed record is read (step S3). If the value l of each attribute A of this record matches the path of the query tree representing the selection condition, the query tree processing unit 522 applies the multiple selection processing having the value l of the attribute A to the record group that has been processed. The analysis instruction is executed (step S4). The query tree processing unit 522 stores the analysis result in an appropriate memory area in association with the path.

  On the other hand, when there is no unprocessed record in the process of step S2 (step S2, No), the query tree processing unit 522 ends the series of multiple selection processes. In this way, the query tree processing unit 522 performs multiple selection processing for all the records in the table D.

  FIG. 10 is a flowchart showing the deviation determination processing procedure. The flowchart of FIG. 10 shows, for example, the timing at which a data analyzer inputs a divergence determination instruction via the input unit 2 before the above-described multiple selection processing is completed for all records retrieved from the table D. It starts with.

  The query tree processing unit 522 first reads a histogram for the entire data (step S11). Specifically, a histogram indicating the distribution of the value 1 of the attribute A subject to the analysis instruction q and a histogram indicating the distribution of the value l of the attribute A specified by the selection condition of each partial data are read.

  Next, the query tree processing unit 522 derives a histogram of the value l of each attribute A of the record group D2 in which the multiple selection process has not been processed (step S12). Specifically, the query tree processing unit 522 first creates a histogram indicating the distribution of the value l of each attribute A of the record group D1 that has been subjected to the multiple selection process for each record. Then, the histogram of the value l of each attribute A of the unprocessed record group D2 is derived by subtracting the histogram of the processed record group D1 from the histogram of the entire data.

  Next, the query tree processing unit 522 uses the histogram of the value 1 of each attribute A of the record group D2 in which the multiple selection processing for each record is not processed, and the analysis result of the partial data that matches each selection condition and the entire data The upper limit value and the lower limit value of the degree of deviation from the analysis result are calculated. Then, for a selection condition where the degree of divergence cannot be maximized, the multiple selection process for subsequent records is skipped (step S13).

  As described above, in the analysis device 1 according to the present embodiment, the multiple query generation unit 51 accepts designation of a plurality of selection conditions having different attributes, and generates a query for each selection condition. In addition, the query tree processing unit 522 determines, for each record stored in the storage unit 4, whether or not each of the selection conditions corresponding to the plurality of queries generated by the multiple query generation unit 51 matches, and for each selection condition In addition, a multiple selection process for analyzing records that meet the selection condition is performed.

  As described above, the query tree processing unit 522 executes the analysis instruction q after executing a plurality of selection processing queries specified by different selection conditions of the plurality of attributes A for each record of the table D at a time. An analysis result of partial data that matches the selection condition is obtained. As a result, the number of accesses to the table D can be reduced to speed up the analysis process.

  The query tree processing unit 522 also analyzes the partial data analysis result and the overall data D analysis result that match each selection condition before the multiple selection processing for each record is completed for all the records in the table D. The degree of divergence is calculated, and a divergence determination process for canceling a plurality of selection processes for subsequent records is performed for a selection condition in which the degree of divergence cannot be maximized.

  When calculating the degree of divergence, the query tree processing unit 522 uses the statistical information of the entire data D and each selection condition acquired before the multiple selection processing for each record is completed for all the records in the table D. The statistical information of the unprocessed record D2 that matches each selection condition is derived using the statistical information of the matched processed record D1, and the degree of divergence is calculated using the statistical information of the unprocessed record D2. To do.

  In this way, the query tree processing unit 522 determines a selection condition in which the degree of divergence between the analysis result of partial data that matches each selection condition and the analysis result of the entire data cannot be maximized. Cancel the multiple selection process of the record. Thereby, it is possible to speed up the analysis process and search for a selection condition for selecting partial data having a high degree of deviation from the analysis result of the entire data.

  As described above, according to the analysis device 1 of the present embodiment, when the attribute A of each selection condition specified in a plurality of selection processes for acquiring partial data from the table D in which data to be analyzed is stored is different. In addition, the number of accesses to the table D can be reduced to speed up the analysis process.

[Other Embodiments]
[program]
It is also possible to create a program in which processing executed by the analysis apparatus 1 according to the above embodiment is described in a language that can be executed by a computer. In this case, the same effect as the above-described embodiment can be obtained by the computer executing the program. Furthermore, the program similar to the above-described embodiment may be realized by recording the program on a computer-readable recording medium, and reading and executing the program recorded on the recording medium. Hereinafter, an example of a computer that executes an analysis program that realizes the same function as the analysis apparatus 1 will be described.

  As shown in FIG. 11, a computer 1000 that executes an analysis program includes, for example, a memory 1010, a CPU 1020, a hard disk drive interface 1030, a disk drive interface 1040, a serial port interface 1050, a video adapter 1060, and a network interface. 1070. These units are connected by a bus 1080.

  The memory 1010 includes a ROM (Read Only Memory) 1011 and a RAM 1012. The ROM 1011 stores a boot program such as BIOS (Basic Input Output System). The hard disk drive interface 1030 is connected to the hard disk drive 1031. The disk drive interface 1040 is connected to the disk drive 1041. For example, a removable storage medium such as a magnetic disk or an optical disk is inserted into the disk drive 1041. For example, a mouse 1051 and a keyboard 1052 are connected to the serial port interface 1050. For example, a display 1061 is connected to the video adapter 1060.

  Here, as shown in FIG. 11, the hard disk drive 1031 stores, for example, an OS 1091, an application program 1092, a program module 1093, and program data 1094. Each table described in the above embodiment is stored in the hard disk drive 1031 or the memory 1010, for example.

  Further, the analysis program is stored in the hard disk drive 1031 as a program module 1093 in which a command executed by the computer 1000 is described, for example. Specifically, a program module describing each process executed by the analysis apparatus 1 described in the above embodiment is stored in the hard disk drive 1031.

  Data used for information processing by the analysis program is stored as program data 1094 in, for example, the hard disk drive 1031. Then, the CPU 1020 reads the program module 1093 and the program data 1094 stored in the hard disk drive 1031 to the RAM 1012 as necessary, and executes the above-described procedures.

  Note that the program module 1093 and the program data 1094 related to the analysis program are not limited to being stored in the hard disk drive 1031, but are stored in, for example, a removable storage medium and read out by the CPU 1020 via the disk drive 1041 or the like. May be. Alternatively, the program module 1093 and the program data 1094 related to the analysis program are stored in another computer connected via a network such as a LAN (Local Area Network) or a WAN (Wide Area Network), and are transmitted via the network interface 1070. It may be read by the CPU 1020.

  As mentioned above, although embodiment which applied the invention made | formed by this inventor was described, this invention is not limited with the description and drawing which make a part of indication of this invention by this embodiment. That is, other embodiments, examples, operational techniques, and the like made by those skilled in the art based on this embodiment are all included in the scope of the present invention.

DESCRIPTION OF SYMBOLS 1 Analyzer 2 Input part 3 Output part 4 Storage part 5 Control part 51 Multiple query production | generation part 52 Multiple query execution part 521 Query tree construction part 522 Query tree processing part

Claims (6)

Receiving a specification of a plurality of selection conditions having different attributes, and generating a query for each of the selection conditions;
For each record stored in the storage unit, it is determined whether each selection condition corresponding to a plurality of queries generated by the generation unit is met, and the record that matches the selection condition is analyzed for each selection condition An analysis unit to
An analysis apparatus comprising: A construction unit that constructs a query tree having the attribute and the attribute value specified by each selection condition as nodes from the query generated by the generation unit;
The analysis unit determines, for each record stored in the storage unit, whether the query tree selection condition constructed by the construction unit is met, and for each selection condition, a record that meets the selection condition The analyzer according to claim 1, wherein the analyzer is analyzed.   The analysis unit diverges between the analysis result of the record that matches each selection condition and the analysis result of all the records in the storage unit before the processing for each record is completed for all the records in the storage unit. The analysis apparatus according to claim 1, wherein the processing for the subsequent records is stopped for a selection condition in which the degree of deviation cannot be maximized.   The analysis unit, the statistical information of all the records in the storage unit, and the processed records that match each selection condition acquired before the processing for each record is completed for all the records in the storage unit The statistical information of unprocessed records that match each selection condition is derived using the statistical information of the unprocessed records, and the degree of deviation is calculated using the statistical information of the unprocessed records. The analyzer described in 1. An analysis method executed by an analyzer,
A generation step of accepting specification of a plurality of selection conditions having different attributes and generating a query for each of the selection conditions;
For each record stored in the storage unit, it is determined whether each selection condition corresponding to a plurality of queries generated in the generation step is met, and a record that matches the selection condition is analyzed for each selection condition Analysis process to
The analysis method characterized by including. A generation step of accepting specification of a plurality of selection conditions having different attributes and generating a query for each of the selection conditions;
For each record stored in the storage unit, it is determined whether each selection condition corresponding to the plurality of queries generated in the generation step is satisfied, and the record that matches the selection condition is analyzed for each selection condition Analysis steps to
An analysis program to make a computer execute.

Images