JPH05158988A - Data base access method - Google Patents

Abstract

PURPOSE:To facilitate designing a hash function according to the distribution of an access key to rialize a high speed hash access method using the hash function in a data base access processing. CONSTITUTION:In a hash access method which directly obtains the storage page of the access-intended record by an access key for the record, an element showing the range for an element to be distributed close to the consecutive numbers as well as the element showing the distribution close to the consecutive number are chosen among the components of the access key (101). The subtraction method is used for an element close to the consecutive number for each value of the element showing the range, and the access is executed by obtaining the storage page based on the subtraction while referring to the control table storing the parameter of the subtraction for each range (104).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a hash access method used for database access processing.

[0002]

2. Description of the Related Art In a conventional hash access method, a storage page of a record to be accessed is directly obtained from a value of an access key of a record to be accessed by a hash function. FIG. 4 is a diagram for explaining a configuration for realizing the hash access method, where 401 is a record to be accessed, 402 is an access key in the record, and 403 is a record.
Is a hash function, and 404 is a storage page of the record. In FIG. 4, the value of the access key 402 is “12
If the division method for obtaining the remainder of “17” represented by “mod 17” in FIG. 4 is used as the hash function 403 for “3”, the storage page 404 of the record becomes “4”. .. In the hash access method, the processing is performed only once for the access target storage page, so that high-speed access is realized. However, if the records are concentrated on a specific storage page, the target storage page becomes full, and the records cannot be stored. At this time, it is stored in another storage page chained to the storage page obtained by the hash function. At this time, i / o required for access
Requires the storage page obtained by the hash function and the storage page chained to the storage page, which reduces the access speed.

For example, when 7 access records can be stored in one storage page and when 70 records are concentrated in a certain page by the hash function, nine storage pages are required in addition to the target storage page. At this time, access requires up to 10 times of i / o.

In the conventional hash access method, a single hash function is applied to the entire area of the access key distribution, and in order to prevent the above-mentioned deterioration of performance, a hash in which records are evenly distributed over all stored pages. Aiming to use functions, many hash functions have been proposed. An example of such a hash function is the division method used in the example of FIG.

[0005]

However, at present, no hash function has been found that always distributes records evenly for any access key distribution. Therefore, in order to use the hash access method efficiently, it is necessary to design the hash function according to the access key distribution, but a design method considering such an access key distribution has not been established. Therefore, the work of designing the hash function is a difficult problem in using the hash access method.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a technique capable of designing an effective hash function in accordance with an access key distribution. is there.

Another object of the present invention is to provide a hash access method using a hash function designed based on the above hash function design method.

The above and other objects and novel features of the present invention will become apparent from the description of this specification and the accompanying drawings.

[0009]

In order to achieve the above object, the present invention provides a hash access method that directly obtains a storage page of an access target record by the access key of the access target record in database access. From the constituent elements, select the element that shows the range in which a certain element has a distribution close to the consecutive number and the element that shows the distribution close to the consecutive number. The division method is used for elements showing a close distribution, and the parameters of the division method for each range are stored in the control table.

Further, at the time of access processing, the most main feature is that the access is executed by obtaining the storage page by the division method while referring to the control table storing the parameters of the division method for each value of the element indicating the range. And

[0011]

According to the above-mentioned means, in the access processing of the database, the hash function can be easily designed according to the distribution of the access key, and the high-speed hash access method using the hash function can be realized.

[0012]

Embodiments of the present invention will be described in detail below with reference to the drawings. FIG. 1 is an explanatory diagram showing a processing procedure for explaining a design method of a hash function used in the hash access method of the present invention.

In FIG. 1, reference numeral 101 denotes a process for selecting an element showing a range in which a certain element has a distribution close to a continuous number and an element close to the continuous number (partitioned element Key element selection processing). Hereinafter, for convenience of explanation, an element indicating a range in which a certain element has a distribution close to a continuous number is referred to as a “section element”, and an element having a distribution close to the continuous number is referred to as a “key element”.

Reference numeral 102 denotes an access key composing column reference used when the processing 101 refers to the column providing access key. The values of the data in the database are determined according to a certain standard. This standard is called "granting standard". For example, a telephone number is an area code indicating a wide area, an area code indicating a small area in the area code, and 4 in the area code.
It consists of digit numbers.

103 is a table showing the digit positions of the partition element and the key element in the access key as a result of the process 101, which is input to the next process 104.

Reference numeral 104 denotes a process (divisor / start storage page determination process) for determining a divisor that is a parameter of the division method and a start storage page indicating a range of storage pages to be used for each value of the partition element. Reference numeral 105 is an actual record for referring to the actual record in the process 104. 10
6 is a list of divisors and starting storage pages for each possible value of the partition element, which is the result of this design method.

The processes 101 and 104 will be described in detail below. FIG. 2 is an explanatory diagram for explaining the details of the process 101 shown in FIG. The process 101 is
As shown in FIG. 2, it comprises processes 201 and 202, and the process 201 is a process (partition element candidate / key element selection process) for selecting (extracting) a candidate for a partition element and a key element for the candidate from the constituent elements. The process 202 is a process of selecting a partition element and a key element (section element / key element selection processing). Reference numeral 203 is a table of assignment criteria of columns forming the access key, 204 is an access key in the access target record in the process 201, and 205 is a process 20.
1 is a component of the access key, 206 is process 2
It is a candidate for the division element in 01. 207 is process 2
12 is a table of the elements (key elements) that are close to the sequential element number corresponding to the segmental element candidate that is the result of 01.

The contents of the process 201 are as follows. ． The access key is decomposed into several components based on the assignment criteria of the columns that the access key comprises. In this example, the assignment criterion 2 of the column that the access key comprises
From 03, the access key 204 is decomposed into three elements C1, C2 and C3 of the access key component 205.

.. When the value of a certain element is fixed and the distribution of another element approaches the distribution of consecutive numbers, the original element is set as a candidate for the segmented element. In the example, when the value of C1 is fixed, C2, C3, and combinations thereof approach a distribution of consecutive numbers. If the value of C2 is fixed, C3 approaches a continuous number distribution. Further, when the combination of C1C2 is fixed, C3 approaches the serial number. As a result, the candidates for the partition element are C1, C2 and C1C2 of the combination thereof.

In process 202, a key element is selected from the candidates selected in process 201, and at the same time, a partition element is selected. In FIG. 2, reference numeral 208 is a list of the digit positions and the number of digits in the access key of the partition element and the key element, which are the results of the processing.

The contents of the process 202 are as follows. ． In the table 207, element candidates having a distribution that is the closest to the continuous number are set as segmented elements. Also, an element showing a distribution close to the serial number at that time is used as a key element. In this example, C3 for the combination of C1 and C2 is the closest to the sequential number, so the partition element is C1C2 and the key element is C.
It becomes 3.

.. Extract the digit position of the partition element and key element in the access key.

In FIG. 1, the process 104 is performed based on the partition element and the key element selected in the process 101. In the process 104, all the data of the actual access record or the sample data is referred to, and the divisor of the parameter of the division method and the starting storage page for applying the division method are determined for each value of the partition element. The processing procedure for that is shown below.

[0024]. Find the starting storage page that indicates the range of storage pages to use. In the division method, records are stored in the storage page for the divisor. For example, if the divisor is “11”, the hash function value changes from “0” to “10”, and the record is stored in the 11 storage pages corresponding to each. As a result, the value obtained by adding the divisor to the start storage page corresponding to the value immediately before the target partition element becomes the target start storage page. For example, if the starting storage page corresponding to the value of the preceding partition element is “123” and the divisor is “11”,
Now, the starting storage page for the value of the target partition element is "134".

[0025]. The number of records that can be stored in one storage page, the number of records for each value of the partition element obtained by referring to actual data, and the usage rate in the storage page, and the storage page required for each value of the partition element Find the number.

.. Calculate the divisor of the division parameter from the required number of stored pages. Here, there are several ways to obtain the divisor. For example, "Key-to-Address Transform Techniq
ues: A Fundamental Performance Study on Large Exist
ing Formatted Files, VYLum, PSYuen, and M.Dodd, CA
As described in "CM, Vol.14, No.4, 1971", there is a method of selecting a prime number close to a required storage page.

By the above-mentioned processing procedure, for each value taken by the partition element, a list in which the divisor corresponding to the value and the starting storage page are paired is created. This list shows the division parameter for each value of each partition element. In the hash access method of the present invention, the access process is performed with reference to this list, so this list will be referred to as a hash control table.

FIG. 3 is an explanatory diagram for explaining a configuration for realizing a hash access method using the hash control table of this embodiment. In FIG. 3, 301 is an access key, 302 is a digit position of a partition element and a key element, 303 is a process of extracting the partition element and the key element from the access key, 3
04 is a hash function, 305 is a hash control table, 306
Indicates a storage page.

The processing procedure of the hash access method using the hash control table of this embodiment will be described below with reference to FIG.

.. The partition element and the key element are extracted from the access key 301 by referring to the digit positions of the partition element and the key element. In FIG. 3, the partition element “129” and the key element “634” are extracted from “129634”.

.. The entry of the hash control table is referred to by the value of the partition element, and the divisor and starting storage page for that value are extracted. In FIG. 3, the divisor from the entry of the partition element “129” in the hash control table is “71”, and the starting storage page is “100”.

.. The storage page is obtained by performing division on the key element and then taking the sum with the start storage page. In FIG. 3, "mod" represents a division method, "63" is obtained from the remainder of "71" of "631", and the sum with the start storage page "100" is obtained, and the access target storage page is It becomes “163”.

As can be seen from the above description, according to this embodiment, as a result of using two kinds of access keys used in an actual system and an access key created by random number generation, comparison with the conventional hash access method is made. Then, the number of records that cannot be stored in the storage page obtained by the hash function is reduced to a maximum of 1/15. Moreover, the performance did not deteriorate as compared with the conventional method. As a result, it became possible to easily design a hash function suitable for the distribution of access keys, which was difficult with the conventional method, and high-speed database access processing became possible.

Although the present invention has been specifically described based on the embodiments, the present invention is not limited to the above embodiments, and various modifications can be made without departing from the scope of the invention. Absent.

[0035]

As described above, according to the present invention, in the access processing of the database, the hash function can be easily designed according to the distribution of the access key, and the high-speed hash access using the hash function is performed. The method can be realized.

[Brief description of drawings]

FIG. 1 is an explanatory view showing a processing procedure for explaining a method of designing a hash function related to a hash access method of the present invention,

FIG. 2 is an explanatory diagram for explaining details of the partition element / key element selection processing shown in FIG.

FIG. 3 is an explanatory diagram for explaining a configuration for implementing a hash access method using the hash control table of the present embodiment,

FIG. 4 is a diagram for explaining a configuration for implementing a hash access method.

[Explanation of symbols]

101 ... Partition element / key element selection process, 102 ... Access key configuration column assignment standard, 103 ... Table showing digit positions of partition element / key element, 104 ... Divisor / start storage page determination process, 105 ... Actual record, 106 ... Divisor / start storage page list (hash control table), 201 ... Partition element candidate / key element selection processing, 202 ... Partition element / key element selection processing, 203 ... Column assignment reference table, 204 ... Access in access target record Key 205, access key component 206, segment candidate, 207
Category element candidate / key element list, 208 ... Category element /
List of digit position and number of digits in access key of key element, 3
01 ... Access key, 302 ... Digit position of partition element / key element, 303 ... Partition element / key element extraction process, 304 ...
Hash function, 305 ... Hash control table, 306 ... Storage page.

Claims (1)

[Claims] 1. In a hash access method in which a storage page of an access target record is directly obtained by an access key of an access target record in database access, an element among access key components has a distribution close to a continuous number. Select an element showing a range and an element showing a distribution close to its continuous number, and for each value of the element showing a range, use the division method for the element showing a distribution close to a continuous number within that range, and A database access method characterized in that access is executed by obtaining a storage page by the division method while referring to a control table storing the parameters of the division method.