JPH10143516A - Database management system, database search method, and recording medium - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To provide a database management system, a database search method, and a database search method capable of obtaining an optimum search order of search conditions by a method that can be executed at any time without burdening a computer. Provision of recording media to be used. SOLUTION: A database management system for searching a database having a plurality of attributes, each of which has a data group having one or a plurality of classification items, according to a predetermined search condition. Comparing means (S26) for comparing the degree of dispersion of the data amount for each attribute to each classification item; and comparing means (S2)
Based on the comparison result of (6), setting means for setting a search order when searching for a plurality of classification items having different attributes as search conditions so as to give a priority search order to the classification items of attributes having a large degree of dispersion (S27). ) And setting means (S2
Search means (S) for searching according to the search order set by (7)
28).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a database management system, a database search method, and a database management method for searching a database consisting of a data group having a plurality of attributes and each attribute having one or a plurality of classification items under a predetermined search condition. It relates to a recording medium used for the implementation.

[0002]

2. Description of the Related Art Assuming a database storing 500,000 personal information of a certain city, whose attributes are gender, occupation and address, for example, by a search condition of "a man and a jewelry occupation," Suppose you search that database. In this case, when the search is performed from the search condition of the classification item “male” whose attribute is gender, as shown in FIG. 7A, the data amount corresponding to the classification item “male” or “female” is about Since it is considered to be 250,000, first, data of about 250,000 "male" is read out, and a search is performed from the data of about 250,000 using the classification item of "jewelry processing" as a search condition. It takes a long time.

On the other hand, when a search is performed from the search condition of the classification item “jewelry processing” whose attribute is “occupation”, as shown in FIG. 7B, the data amount corresponding to the classification item “jewelry processing” is small (large). Therefore, first, the data of several tens is read, and a search may be performed from the data of several tens using the classification item of “male” as a search condition. Ends with That is, when searching the database using a plurality of search conditions, the required time may greatly differ depending on the order of the search conditions. This is because some of the search conditions have a large corresponding data amount (the sum of the data amounts for each attribute is the same (500,000) for each attribute), and the search is performed from the classification item. Occurs in the case.

[0004]

In such a case, in a conventional method, for example, in a multiple condition simultaneous search method, a plurality of searches are simultaneously performed in order to effectively use an idle portion of computer capability. However, when a plurality of users access the computer at the same time and there is no idle portion, this is not an efficient method. Further, in a method of creating a data amount distribution table for each search condition and performing a search from search conditions with a small data amount, it is necessary to count all data in the database for each search condition, which is a huge amount of time for creating a distribution table. It is not effective because it takes time and cannot be executed at any time.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned circumstances, and provides a database capable of performing a search without burdening a computer and obtaining an optimum search order of search conditions by a method that can be executed at any time. It is an object of the present invention to provide a management system, a database search method, and a recording medium used for executing the method.

[0006]

According to a first aspect of the present invention, there is provided a database management system having a plurality of attributes, each of which has one or more classification items. A database management system for setting a search order when a plurality of classification items having different attributes are searched as search conditions, and a search unit for searching according to the search order set by the setting unit. It is characterized by.

[0007] In the database management system according to the first invention, the setting means sets a search order when a plurality of classification items having different attributes are searched as search conditions. Then, the search means searches according to the set search order. As a result, it is possible to obtain the optimum search order of the search conditions and perform the search in a short time without burdening the computer.

The database management system according to a second aspect of the present invention includes a comparing unit that compares the degree of dispersion of the data amount for each attribute to each classification item, and the setting unit includes:
According to the comparison result of the comparing means, a priority search order is given to the category items of the attribute having the large degree of dispersion.

[0009] A database management system according to a third aspect of the present invention is characterized in that the database management system further comprises a calculating means for calculating the degree of dispersion for each attribute.

According to a fourth aspect of the present invention, in the database management system, the degree of dispersion is assigned to each of the classification items by assigning each term of an arithmetic progression having a predetermined equality to the respective classification items in a predetermined order according to the data amount of each classification item. , And the variance of the frequency distribution for each class is calculated for each attribute, with the data amount being a frequency and the respective terms being a class.

[0011] A database management system according to a fifth invention is characterized in that the variance is a sample variance obtained by an estimation method.

A database search method according to a sixth aspect of the present invention
In a database search method for searching a database composed of a data group having a plurality of attributes and each attribute having one or a plurality of classification items under a predetermined search condition, the data amount for each attribute may be distributed to each classification item. By comparing the magnitudes and setting a search order in the case of searching for a plurality of classification items having different attributes as a search condition so as to give a priority to the classification items of the attribute having a large dispersion degree, The search is performed according to

[0013] In the database management system according to the second invention and the database search method according to the sixth invention, the comparing means compares the degree of dispersion of the data amount for each attribute to each classification item. Then, the setting means sets a search order in a case where a plurality of classification items having different attributes are searched as a search condition based on the comparison result so as to give a priority to the classification items of the attribute having a large degree of dispersion, Search means searches according to the set search order. As a result, the data is divided into a large number of classification items whose attribute is the residential area shown in FIG. 7C, and the difference in the data amount between the classification items is small and the data amount is scattered among the classification items. Can be searched for from an attribute that is not likely to search for a classification item having a large data amount, so that the search can be completed in a short time.

[0014] In the database management system according to the third aspect of the invention, the calculating means calculates the degree of dispersion of the data amount for each attribute to each classification item. Then, the comparison unit compares the degree of dispersion of the data amount for each attribute to each classification item, and the setting unit searches for a plurality of classification items having different attributes as search conditions based on the comparison result. The order is set so as to give priority to the classification items of the attribute having a large degree of dispersion, and the search means performs a search according to the set search order. As a result, the degree of dispersion of the data amount for each attribute to each classification item can be calculated.
The data is divided into a large number of classification items whose attribute is the residential area shown in (c), the difference in the data amount of each classification item is small, and the degree of dispersion of the data amount to each classification item is large. A search can be performed from an attribute that does not have a risk of searching for a classification item having a large data amount, and the search can be completed in a short time.

A database search method according to a seventh aspect of the present invention
The degree of dispersion is assigned to each item of the arithmetic progression having a predetermined equal difference in a predetermined order according to the data amount of each classification item, and the data amount is set as a frequency, and the respective items are classified into classes. The variance of the frequency distribution for each class is calculated and obtained for each attribute.

[0016] In the database management system according to the fourth invention and the database search method according to the seventh invention, the arithmetic means has a predetermined equal difference in each classification item in a predetermined order according to the data amount of each classification item. Each term of the difference sequence is assigned, the data amount is set as a frequency, and each item of the arithmetic sequence is set as a class, and the variance of the frequency distribution for each class is calculated for each attribute. As a result, the degree of dispersion of the data amount for each attribute to each classification item can be calculated, so that the data is divided into a large number of classification items such as those shown in FIG.
The difference in the data amount of each classification item is small, the degree of dispersion of the data amount to each classification item is large, and it is possible to search from an attribute that does not have a risk of searching for a classification item with a large data amount, and search in a short time Can be terminated.

According to an eighth aspect of the present invention, there is provided a database search method comprising:
The variance is a sample variance obtained by an estimation method.

In the database management system according to the fifth invention and the database search method according to the eighth invention, the calculating means randomly extracts the variance, which is the degree of dispersion of the data amount for each attribute to each classification item, and extracts each data. And by estimation
Calculate as sample variance. Then, the comparison unit compares the degree of dispersion of the data amount for each attribute to each classification item, and the setting unit searches for a plurality of classification items having different attributes as search conditions based on the comparison result. The order is set so as to give priority to the classification items of the attribute having a large degree of dispersion, and the search means performs a search according to the set search order. As a result, the degree of dispersion of the data amount for each attribute to each classification item can be calculated in a short time, so that the search can be completed in a short time without imposing a load on the capability of the computer.

According to a ninth aspect of the present invention, there is provided a recording medium storing a computer program for searching a database consisting of a data group having a plurality of attributes and each attribute having one or a plurality of classification items under a predetermined search condition. Comparing the degree of dispersion of the data amount for each attribute to each classification item in a recording medium, and searching for a plurality of classification items having different attributes as search conditions based on the comparison result of the step. Setting a search order so as to give a priority to the classification items of the attribute having a large degree of dispersion, and performing a search in accordance with the search order.

In a computer controlled by a computer program recorded on the recording medium, the data amount of each attribute is compared in the degree of dispersion to each classification item, and a plurality of classifications having different attributes are determined based on the comparison result. The search order in the case where an item is searched as a search condition is set so as to give priority to the classification items of the attribute having a large degree of dispersion, and the search is performed according to the search order. As a result, as shown in FIG.
The data is divided into a large number of classification items, the difference in the data amount of each classification item is small, the degree of dispersion of the data amount to each classification item is large, and there is no risk of searching for a classification item with a large data amount. Therefore, the search can be completed in a short time.

According to a tenth aspect of the present invention, there is provided a recording medium, comprising: assigning, to each of the classification items, respective items of an arithmetic progression having a predetermined equality in a predetermined order according to the data amount of each classification item; , And calculating the variance of the frequency distribution for each class for each of the attributes to obtain the degree of dispersion.

In a computer controlled by a computer program recorded on the recording medium, the arithmetic means comprises: for each classification item, an arithmetic progression having a predetermined equality in a predetermined order according to the data amount of each classification item. Are assigned, the data amount is set as a frequency, and each term of the arithmetic progression is set as a class, and the variance of the frequency distribution for each class is calculated for each attribute to obtain the degree of dispersion. As a result, the degree of dispersion of the data amount for each attribute to each classification item can be calculated.
The data is divided into a large number of classification items whose attributes are residential areas as shown in, and the difference in the data amount of each classification item is small, the degree of distribution of the data amount to each classification item is large, and the data amount is large. A search can be performed from an attribute that does not have a risk of searching for a classification item, and the search can be completed in a short time.

[0023] In a recording medium according to an eleventh aspect, the variance is a sample variance obtained by an estimation method.

In the computer controlled by the computer program stored in the recording medium, the arithmetic means randomly extracts the variance, which is the degree of dispersion of the data amount for each attribute to each classification item, by extracting each data. By estimation,
Calculate as sample variance. Then, the comparison unit compares the degree of dispersion of the data amount for each attribute to each classification item, and the setting unit searches for a plurality of classification items having different attributes as search conditions based on the comparison result. The order is set so as to give priority to the classification items of the attribute having a large degree of dispersion, and the search means performs a search according to the set search order. As a result, the degree of dispersion of the data amount for each attribute to each classification item can be calculated in a short time, so that the search can be completed in a short time without imposing a load on the capability of the computer.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the drawings showing an embodiment. FIG. 1 is a block diagram showing a configuration of an embodiment of a database management system and a database search method according to the present invention. The database management system includes a CPU 10, a ROM 11,
A RAM 12, a database 13 including a magnetic disk and data recorded thereon, and a CR for generating a signal to be displayed on the CRT 15 and controlling the display on the CRT 15
And a T control circuit 14.

This database management system also includes:
A keyboard 16 for a user to input, a printer 17, a disk drive 18 for driving a flexible disk, a CPU 10, a ROM 11, a RAM 1
2, a database 13, a CRT control circuit 14, a keyboard 16, a printer 17, and a bus 17 for interconnecting a disk drive 18.

The disk drive 18 is loaded with a flexible disk FD (see FIG. 6) in which a computer program for searching the database 13 is recorded. This computer program is executed by the CPU 10 in a state where all or a part thereof is read into the RAM 12.
Control.

Hereinafter, a database management system having such a configuration will be described. This database management system calculates the degree of dispersion of the data amount for each attribute to each classification item at a predetermined time. At this time, for example, FIG.
(B) a as shown, the distribution table categories are arranged in random, as shown in FIG. 4 (a), the amount of data y _m is largest in the classification item as a central position, in order large amount of data Each classification item is arranged by sorting from the center position, and each term of the arithmetic progression having a predetermined equal difference in each classification item {x _m-2 , x _m-1 ,
x _m , x _{m + 1} , x _{m + 2} }, x _n (n is the number of classification items of the entire attribute) are sequentially allocated and quantified. Then, using the data amount y as a frequency and each term x as a class, the variance of the frequency distribution for each class x is calculated for each attribute to obtain the degree of dispersion.

Here, when the data amount y is sufficiently large, the frequency distribution for each class x can be regarded as a normal distribution, so that the variance can be obtained by estimating the sample variance. The estimation, as with the data of the certain size of the data volume number Y _t from the data amount y _t of the entire attributes extracted at random were counted for each classification item, shown in FIG. 4 (a) method a manner, as the classification items central position of the data amount most Y _m, in order a large amount of data by distributing the center arranges each classification item, ‥ terms of arithmetic progression having a predetermined arithmetic to each classification item ‥ X _m-2 , X _m-1 , X _m , X _{m + 1} , X
_{m + 2} ‥‥, X _p (p is the number of classification items to which the randomly extracted data belongs) is sequentially assigned, the data amount Y is set as a frequency, each term X is set as a class, and the sample variance of the frequency distribution for each class X is calculated. What is necessary is just to calculate.

In this database management system, the degree of dispersion of the data amount for each attribute to each classification item is calculated. Then, the degree of dispersion of the data amount for each attribute to each classification item is compared, and the search order when searching for a plurality of classification items having different attributes as a search condition is determined by the classification item of the attribute having a large degree of dispersion. Is set so as to give a priority, and a search is performed according to the set search order.

FIG. 2 is a flowchart showing the operation of the database management system for calculating the degree of dispersion of the data amount for each attribute to each classification item. For example, when the database is updated (S10), the database management system confirms that there is no access to the database (S12). Then, the data of a predetermined size of data amount number Y _t determined from the data amount y _t of the entire attribute randomly selected and counted for each classification item (S1
4).

Next, this database management system:
The categories were counted (S14) data of each classification item most Y _m as a central position, in the same manner as in the method shown in FIG. 4 (a), the amount of data _{‥‥ Y m-2, Y m} -1 , Y _m , Y
_{m + 1, Y m + 2} ‥‥, Y p arrange each classification item sorting from the center position in order, the more terms ‥‥ X _m-2 of arithmetic progression having a predetermined arithmetic to each classification item, X _m−1 , X _m , X _{m + 1} ,
X _{m + 2} }, X _p (p is the number of classification items to which the randomly extracted data belongs) are sequentially assigned (S16).

Next, this database management system:
The randomized data is quantified belonging category _{‥‥ X m-2, X m} -1, X m, X m + 1, X m + 2 ‥‥, X p
Calculating the average value X calculated by the following expression _a X _a of _{= (ΣX i · Y i)} / Y t (i = 1~p) (1).

Next, using the average value X _a calculated in (1), the sample variance s ² of the frequency distribution for each class X is calculated as follows: s ² = (Σ (X _i −X _a ) ² · Y _i ) / Y _t (i = 1 to p) (2) to calculate the sample variance s ² as the data amount y of the attribute.
(S1)
8). The database management system calculates the sample variance s ² for all attributes (S 19) by the above-described method, and determines the data amount y of the attribute to be scattered to the classification item (class x) (S 18).

FIG. 3 is a flowchart showing a search operation of the database management system. When the user selects a data search function using the keyboard 16 while viewing the CRT 15 (S20), the database management system displays a data search screen on the CRT 15 (S22). The user uses the keyboard 16 to input a plurality of classification items having different attributes (for example, “I am a man and work in jewelry processing” (see FIG. 7)) as search conditions while viewing this data search screen (S24). .

When the data item to be searched is input (S24), the database management system compares the sample variances for each attribute ("sex", "occupation") to which the classification item as the search condition belongs (S26). The search order of the classification item (“jewelry processing”) of the attribute (“occupation”) with the largest sample variance is the first, and the attribute with the second largest sample variance (in this case, the sample variance is small) The search order of “sex” is set as the second (S27).

Next, in accordance with the set search order (S27), a search is performed according to the classification item ("jewelry processing") of the attribute ("occupation") having the first search order. Then, a search is performed according to the classification item ("male") of the attribute ("sex") having the second search order (S28). Next, the searched data is displayed on the CRT 15 (S3).
0).

In this case, as shown in FIG. 7A, the (sample) variance of the frequency (data amount) distribution table for each classification item in which a large amount of data is concentrated on a small number of classification items is as shown in FIG. )
As shown in FIG. As shown in FIG. 5B, a large amount of data is divided into many classification items, and there is a difference between the data amounts. The (sample) variance of the frequency distribution table for each classification item is shown in FIG. It has a medium value as shown. (C)
As shown in FIG. 5A, the (sample) variance of the frequency distribution table for each classification item, in which a large amount of data is divided into many classification items and there is no significant difference in the respective data amounts, is as shown in FIG. Become larger.

FIG. 6 is an explanatory diagram for explaining an embodiment of a recording medium according to the present invention. In this embodiment, at a predetermined time, for each attribute, each term of the arithmetic progression having a predetermined arithmetic difference is assigned to each classification item in a predetermined order according to the data amount of each classification item (S2). And a step (S4) of calculating the sample variance of the frequency distribution of each class for each attribute to set the degree of dispersion by using the data amount as a frequency and each term of the arithmetic progression as a class, and a plurality of classification items having different attributes Is used as a search condition, a step (S6) of comparing the degree of dispersion of the data amount for each attribute to each classification item, and a plurality of classification items having different attributes are searched as a search condition based on the comparison result. Setting the search order in such a case that priority is given to the classification items of the attributes having a large degree of dispersion (S7); and setting the search order of the attributes having a large degree of dispersion in accordance with the set search order. Flexible disk FD that are recorded thereon a computer program PR and a step (S8) for retrieving from the item has been loaded into the disk drive DD.

The computer program PR is read by the computer CP and controls the computer CP.
Other configurations and operations are the same as those of the above-described embodiment of the database management system and the database search method, and thus the description is omitted.

[0041] In the embodiment described above, although the average value X _a classification item randomly extracted data is quantified belongs, the data amount by calculating the largest number of categories X _m as an average value Can obtain substantially the same result. In this case, the classification items and the quantity as X _m = 0 integer ‥‥ -
2, -1, 0, +1, +2}, the operation can be simplified. In the above-described embodiment, the (sample) variance s ² is used as the degree of dispersion. However, the same applies when the (sample) standard deviation s is used as the degree of dispersion.

Further, as shown in FIG. 4B, the classification items are arranged in descending order of the data amount, the classification item having the largest data amount X _m = 0, and the quantified classification items are represented by an integer ‥‥ 0,
1, 2, 3, 4}, the degree of dispersion of the data amount for each attribute to each classification item can be obtained even if the calculation is performed by the above-described method, and the same result as in the above-described embodiment can be obtained. Obtainable.

[0043]

According to the database management system according to the first aspect of the present invention,
It is possible to obtain the optimum search order of the search condition and perform the search in a short time without putting a burden on the computer.

According to the database management system according to the second invention and the database search method according to the sixth invention, data is divided into a large number of classification items, and the difference in data amount between the classification items is small, so that each classification item has a data amount. Can be searched from an attribute that has a high degree of dispersion and that does not have a risk of searching for a classification item with a large data amount, so that the search can be completed in a short time.

According to the database management system according to the third and fourth inventions and the database search method according to the seventh invention, it is possible to calculate the degree of dispersion of the data amount for each attribute to each classification item. , The difference between the data amounts of the respective classification items is small, the degree of dispersion of the data amount among the respective classification items is large, and it is possible to search from an attribute that does not have a risk of searching for the classification item with a large data amount. The search can be completed in time.

According to the database management system according to the fifth invention and the database search method according to the eighth invention, the degree of dispersion of the data amount for each attribute to each classification item can be calculated in a short time. The search can be completed without burden and in a short time.

According to the computer controlled by the computer program in which the recording medium according to the ninth aspect of the invention is recorded, the data is divided into a large number of classification items, the difference in the data amount of each classification item is small, and Since the degree of dispersion to the items is large, and a classification item having a large data amount is not likely to be searched, it is possible to search from an attribute, so that the search can be completed in a short time.

According to the computer controlled by the computer program recorded on the recording medium according to the tenth aspect, the degree of dispersion of the data amount for each attribute to each classification item can be calculated. Are separated, the difference in the data amount of each classification item is small, the degree of dispersion of the data amount to each classification item is large, and it is possible to search from an attribute that does not have a risk of searching for a classification item with a large data amount. To end the search.

According to the computer controlled by the computer program recorded on the recording medium according to the eleventh aspect, the degree of dispersion of the data amount for each attribute to each classification item can be calculated in a short time. The search can be completed in a short time without putting a burden on the ability.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of a database management system and a database search method according to the present invention.

FIG. 2 is a flowchart showing the operation of the database management system according to the present invention for calculating the degree of dispersion of the data amount for each attribute to each classification item.

FIG. 3 is a flowchart showing a search operation of the database management system according to the present invention.

FIG. 4 is an explanatory diagram for describing a method of calculating the degree of dispersion of the data amount for each attribute to each classification item as a variance.

FIG. 5 is an explanatory diagram for explaining the magnitude of dispersion.

FIG. 6 is an explanatory diagram illustrating an embodiment of a recording medium according to the present invention.

FIG. 7 is a data amount distribution diagram for explaining the relationship among data attributes, classification items, and data amount in a database.

[Explanation of symbols]

 10 CPU 12 RAM 13 Database 15 CRT 18, DD Disk Drive CP Computer FD Flexible Disk PR Computer Program

Claims (11)

[Claims] 1. A database management system for searching a database consisting of a data group having a plurality of attributes and each attribute having one or a plurality of classification items according to a predetermined search condition, comprising: A database management system comprising: a setting unit that sets a search order when searching as a search condition; and a search unit that searches according to the search order set by the setting unit. 2. The method according to claim 1, further comprising: comparing means for comparing the degree of dispersion of the data amount for each attribute to each classification item, wherein the setting means determines that the degree of dispersion is large based on a comparison result of the comparing means. 2. The database management system according to claim 1, wherein a priority search order is assigned to attribute classification items. 3. The database management system according to claim 2, further comprising a calculation unit that calculates the degree of dispersion for each attribute. 4. The scatter degree is assigned to each of the classification items, in a predetermined order according to the data amount of each classification item, each term of an arithmetic progression having a predetermined equal difference, and the data amount is a frequency, 4. The database management system according to claim 2, wherein the variance of the frequency distribution for each class is calculated for each attribute, with each of the terms as a class. 5. 5. The database management system according to claim 4, wherein the variance is a sample variance obtained by an estimation method. 6. A database search method for searching a database consisting of a data group having a plurality of attributes and each attribute having one or a plurality of classification items under a predetermined search condition, wherein each classification item of the data amount for each attribute The degree of dispersion is compared, and a search order when searching for a plurality of classification items having different attributes as a search condition is set so as to give a priority to the classification items of the attribute having the higher degree of dispersion. And a search according to the search order. 7. The degree of scatter is assigned to each of the classification items, in a predetermined order according to the data amount of each classification item, each term of an arithmetic progression having a predetermined arithmetic difference, and the data amount is a frequency, 7. The database search method according to claim 6, wherein each of the terms is set as a class and the variance of the frequency distribution for each class is calculated for each attribute. 8. The database search method according to claim 7, wherein the variance is a sample variance obtained by an estimation method. 9. A recording medium storing a computer program for searching a database having a plurality of attributes, each of which has a data group having one or a plurality of classification items, under a predetermined search condition, Comparing the degree of dispersion of the data amount for each attribute to each classification item; and, based on the comparison result of the step, determining the search order when a plurality of classification items having different attributes are searched as search conditions. A recording medium comprising: a step of setting a priority order to a category item of the attribute having a large degree; and a step of searching according to the search order. 10. A step of allocating each term of an arithmetic progression having a predetermined equality to each of the classification items in a predetermined order according to the data amount of each classification item; 10. The recording medium according to claim 9, further comprising: calculating a variance of the frequency distribution for each class for each of the attributes to obtain the degree of dispersion. 11. The recording medium according to claim 10, wherein the variance is a sample variance obtained by an estimation method.