JP2004145729A - Data management device and program and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a data management device which can search a data-registered packet at high speed without being affected by the size of a hash table or the number of data stored in the hash table when taking out all the data registered in the hash table. <P>SOLUTION: The data management device is provided with a hash table storage means 14 consisting of: the hash table; a chain list of a component holding data information starting from the packet obtained by a data hash value to be registered and located on the hash table; and link information holding the association of the registered packet on the hash table. In registering the data information on the chain list and in registering the data information on a new packet on the hash table, the link information is modified by assigning the new packet as a coming packet following the registered packet. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a data management device, a program, and a recording medium for registering and retrieving data, and more particularly to a data management device for obtaining a data storage location using a hash function.
[0002]
[Prior art]
One of techniques for performing data management such as data registration and search in a computer is a hash method (see Non-Patent Document 1 and Patent Document 1).
Since the hash method compresses data into a hash value of a certain length, a collision may occur in the hash value.
For example, as shown in FIG. 13, when the size of the hash table is n, the hash table is divided into n buckets, and each bucket is indicated by an index from 1 to n.
In the case of FIG. 13, the hash values collide at the index n, and are solved by using the chain method.
[0003]
When registering certain data using this hash method, the following procedure is used.
(1) A hash function is applied to data to be registered to obtain a hash value.
(2) Obtain a bucket using the obtained hash value as an index.
(3) Check if there is any registered in the obtained bucket.
(4) Like the bucket of index 2 shown in FIG. 13, zero of the empty pointer is stored, and data to be registered (or a pointer to the data) is stored in a bucket in which nothing is registered. Generate an element and register a pointer to that element in the bucket (see index 1 or 3).
(5) Alternatively, the data to be registered (or the pointer to the data) is stored in the bucket in which the pointer to the head element of the chain is already stored, such as the buckets of index 1 and index 3 shown in FIG. Generate the stored element and register a pointer to that element in the bucket. In the generated element, a pointer to the original head element is stored (see index n). As a result, the newly generated element becomes the head of the chain, and the original head element becomes the second element from the head of the chain.
[0004]
When a data value is registered using a hash table as described above, data for which the same hash value is obtained is registered in a chain list stored in a bucket using the hash value as an index.
[0005]
To retrieve the data registered as described above, the following procedure is performed.
(1) The index of the bucket currently being processed is initialized to 1.
(2) If the index of the bucket currently being processed is greater than n, all buckets have been processed, and the data fetching process ends.
(3) The pointer to the head element of the chain list stored in the bucket currently being processed is extracted and used as the pointer to the element currently being processed.
(A) If the pointer to the element currently being processed is an empty pointer of zero, all the elements currently being processed have been processed, and the process proceeds to (4).
(B) Otherwise, retrieve the data stored in the element pointed to by the pointer.
(C) The pointer to the next element stored in the element indicated by the pointer is taken out, set as the pointer to the element currently being processed, and the process returns to (A).
(4) Add 1 to the index of the bucket currently being processed, and return to (2).
[0006]
In the “table management method” of Patent Document 2, in order to enable all elements registered in a hash table to be traced in an independent order that does not depend on the registration order, the following elements are obtained for each element. Information (for example, a pointer) is stored. By using this information, regardless of the order in which the data is registered in the hash table, all the elements are arranged in an arbitrary order, for example, in an order in which the data held in the elements are arranged in ascending order. be able to.
[0007]
[Non-patent document 1]
Hagiwara Hiroshi and Nishihara Seiichi, "Modern Data Structure and Programming Techniques",
Ohmsha, October 1987
[Patent Document 1]
Japanese Patent No. 3056704 (prior art)
[Patent Document 2]
JP 05-081022 A
[0008]
[Problems to be solved by the invention]
As described above, when fetching all data registered in the hash table, if a pointer to the head element of the chain is not stored in a certain bucket, the head of the chain is set after the bucket of (index + 1) of that bucket. It is necessary to check whether pointers to elements are registered one by one.
[0009]
This study shows that as the size of the hash table increases, and as the number of data stored in the hash table decreases, the pointer to the head element of the chain is often not registered in the bucket. The time required to search for the bucket required to obtain the data of the data becomes longer.
[0010]
The technique of Patent Literature 2 is for solving this problem. Compared with a hash table using a general chain method, only the elements are traversed one after another. Can be taken out at high speed.
However, since information for obtaining the next element is stored for each element, an area for the information is required in proportion to the number of registered elements.
[0011]
The present invention has been made in consideration of the above-described circumstances, and when extracting all data registered in the hash table, the size of the hash table and the number of data stored in the hash table are not affected. It is an object of the present invention to provide a data management device, a program, and a recording medium capable of quickly searching for a bucket in which data is registered.
[0012]
[Means for Solving the Problems]
In order to solve the above-mentioned problem, a data management apparatus according to claim 1 of the present invention is a data management apparatus that determines a storage location of data to be registered using a hash function. A chain list in which elements holding information on data starting from a bucket on the hash table obtained by applying the above are linked by a chain, and link information holding an association between registered buckets on the hash table. A hash table storage unit, and a data registration unit that registers information about the data in a chain list indicated by a bucket obtained by applying a hash function to the data to be registered and stores the information in the hash table storage unit, When registering to a new bucket on the hash table, Wherein the data of a new bucket is possible to change the link information as a bucket to follow the next registered buckets.
[0013]
According to a second aspect of the present invention, in the data management device according to the first aspect, the link information includes information indicating a position of a bucket to be followed before a bucket in which data has been registered, and data in which the data has been registered. And information indicating the position of the bucket to be traced after the bucket.
According to a third aspect of the present invention, in the data management device according to the first or second aspect, the link information is an index indicating a bucket.
[0014]
According to a fourth aspect of the present invention, in the data management apparatus according to the first, second, or third aspect, the data management apparatus further includes a data search unit that searches for registered data. A bucket in which data is registered in the hash table is extracted from the link information, and registered data is extracted from a chain list linked from the bucket.
[0015]
According to a fifth aspect of the present invention, in the data management apparatus according to the first, second, third, or fourth aspect, the data management apparatus further comprises a data deletion unit for deleting the registered data, wherein the data deletion unit performs the designated deletion. Data is deleted from the chain list, and when there is no more data registered in the chain list, a bucket indicating the deleted data is deleted from the link information.
[0016]
A program according to a sixth aspect of the present invention is a program for causing a computer to execute the functions of the data management device according to any one of the first to fifth aspects. A recording medium according to a seventh aspect of the present invention is a computer-readable recording medium storing the program according to the sixth aspect.
[0017]
Therefore, when retrieving all data registered in the hash table, the bucket in which the data is registered can be searched at high speed without affecting the size of the hash table or the number of data stored in the hash table. .
[0018]
Furthermore, the bucket can be searched at high speed regardless of the size of the hash table or the number of data stored in the hash table, regardless of whether the data is retrieved in the normal or reverse order.
[0019]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, an embodiment of a data management device according to the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of a computer system constituting an embodiment of a data management device according to the present invention.
In FIG. 1, this embodiment includes a server 10 and a terminal 20.
[0020]
The terminal 20 allows the user of the system to specify the data itself to be registered / searched / deleted with respect to the data management device executed on the server 10 and the data file in which the data is recorded, and to execute the obtained operation. A computer used to output results.
In general, the terminal 20 is provided with a pointing device such as a keyboard and a mouse for inputting an inquiry request, and a display for displaying an inquiry result.
The terminal 20 includes a CPU, a memory, and a hard disk, and can execute a program.
[0021]
The server 10 is a computer that is connected to the terminal 20 via the communication network 30, performs a process according to an operation request from the terminal 20, and returns a result to the terminal 20.
The server 10 includes a CPU, a memory, and a hard disk, and can execute a program for data management.
Data managed by the server 10 is recorded on a recording medium such as a hard disk.
[0022]
Here, the communication network 30 is a transmission path for coupling between the server 10 and the terminal 20, and is generally realized by a cable, and TCP / IP is used as a communication protocol. However, the transmission path is not limited to a cable, but may be a wired or wireless communication as long as the communication protocol between them is the same. For example, a LAN (Local Area Network) using a public line or a dedicated line, a WAN ( Wide Area Network, the Internet, and the like can be used.
[0023]
When all the functions of the data management device of the present invention are executed by the terminal 20, the server 10 and the communication network 30 described above are not required.
[0024]
FIG. 2 is a block diagram showing a functional configuration of the data management device according to the present invention. In FIG. 2, the data management device includes a data input / output unit 21, a data registration unit 11, a hash table storage unit 14, a data search unit. 12 and data deletion means 13.
[0025]
The data input / output means 21 operates at the terminal 20, receives data registration, search, and deletion requests input by the user via the terminal 20, processes the requests, and outputs the results.
[0026]
The data registration unit 11, the data search unit 12, the data deletion unit 13, and the hash table storage unit 14 operate on the server 10 and constitute a part of the function of the data management device.
[0027]
First, the data structure of the hash table used by the data registration unit 11, the data search unit 12, and the data deletion unit 13, the hash table storage unit 14 that stores the chain list and the link information will be described. This hash table storage means 14 is stored on the memory of the server 10 or on a hard disk.
[0028]
FIG. 3 shows an example of the data structure of the hash table storage means 14 used in the present embodiment. The data structures of the hash table H and the chain list C are the same as those described in the related art.
[0029]
That is, when the size of the hash table H is n, the hash table H is divided into n buckets, and each bucket is indicated by an index from 1 to n.
Each bucket for which no data is registered is set to zero.
When registering data, a hash function is applied to the data, a hash value is obtained, and a bucket having the hash value as an index is obtained.
(1) If there is nothing registered in the obtained bucket (the value stored in the bucket is zero), an element E is generated, and a pointer to the element E is registered in the bucket.
The element E includes data to be registered (or a pointer to the data) Ev and a pointer Ep to another element E of the data having the same hash value.
If the pointer Ep of the element E is set to zero, it indicates that there is no element to be chained next.
[0030]
(2) Alternatively, if there is a bucket registered in the obtained bucket, an element E storing the data to be registered (or a pointer to the data) is generated, and the chain pointer Ep of the generated element E is stored in the chain pointer Ep. Stores a pointer to the head element of the chain list registered in the bucket, and registers a pointer to the generated element E in the bucket.
[0031]
Hereinafter, the present embodiment will be described on the assumption that an index indicating a bucket is used as the expression format of the link information.
In FIG. 3, the link information includes a rear link start index B and rear link information N, and a front link start index A and front link information P.
[0032]
The rear link start index B and the rear link information N are each configured as one array by arranging the element of index 0 and then the elements corresponding to the buckets of indexes 1 to n.
The back link start index B stores the index of the bucket to be searched first when acquiring all the data registered in the hash table H. The post-link information N stores the index of the bucket to start searching next after searching for the chained element from the bucket. When the value of the subsequent link information N is zero, it indicates that there is no next bucket to be searched.
[0033]
For example, in FIG. 4, since the stored value of the post-link start index B (index 0) is 3, the first search is performed from the bucket of index 3.
Since n is stored in the post-link information N of the bucket of the index 3, the bucket searched next to this bucket is the bucket of the index n.
Further, since zero is stored in the post-link information N of the index 1, there is no bucket to be searched next to this bucket.
[0034]
Similarly, the previous link start index A and the previous link information P are configured as one array by arranging the element of the index 0 and then the elements corresponding to the buckets of the indexes 1 to n.
The front link start index A stores the index of the first bucket to be searched when the buckets are arranged in the reverse order of the tracing by the rear link information when acquiring all the data registered in the hash table H. You. The previous link information P stores an index of a bucket to start searching next after searching for an element chained from the bucket. When the value of the previous link information P is zero, it indicates that there is no next bucket to be searched.
[0035]
For example, in FIG. 4, since the stored value of the previous link start index A (index 0) is 1, the first search is performed from the bucket of index 1.
Since n is stored in the preceding link information P of the bucket of index 1, the bucket searched next to this bucket is the bucket of index n.
Since zero is stored in the previous link information P of the index 2, there is no bucket to be searched next to this bucket.
[0036]
In response to a data registration request from the data input / output unit 21, the data registration unit 11 finds a chain in which the given data is to be registered, and registers the data as the first element of the chain.
The processing procedure of the data registration unit 11 will be described with reference to the flowchart shown in FIG.
Here, before registering the data, the values of the buckets of the previous link start index A, the rear link start index B, and the hash table H are initialized to zero.
[0037]
A hash function is applied to the registered data passed from the data input / output unit 21 to obtain a hash value (steps S1 and S2).
A bucket having the obtained hash value as an index is obtained (step S3). Hereinafter, the calculated bucket index is defined as i.
An element for storing data to be registered (or a pointer to the data) is generated (step S3).
This element is in the chain list C and corresponds to the element E in FIG. 3. Data to be registered (or a pointer to the data) is stored in Ev, and zero is stored in Ep.
[0038]
It is checked whether the content of the bucket at index i is zero (step S5).
If the content of the bucket at index i is not zero (YES in step S5), since the pointer to the head element of the chain has already been registered, the content of the bucket at index i is stored in the Ep area of the element of the generated chain list C. Is stored (step S6), and the process proceeds to step S11. As a result, the element just created becomes the head of the chain list.
[0039]
On the other hand, if the content of the bucket at index i is zero (NO at step S5), the previous link information at index i is stored as zero (step S7).
I is stored in the previous link information using the content of the rear link start index as an index (step S8).
The content of the post-link start index is stored as post-link information of index i (step S9).
I is stored in the post-link start index (step S10).
Finally, the pointer to the element generated in the chain list is registered in the bucket at index i (step S11), and the data registration ends.
[0040]
For example, when data with a hash value of 2 is registered in the hash table storage unit 14 in the state of FIG. 4, the data is registered as shown in FIG.
[0041]
If the data search request from the data input / output unit 21 searches for one piece of data, the data search unit 12 obtains an element in the chain list in which the given data is registered from the hash table, and Returns the result to the data input / output means 21 depending on whether or not holds the data to be searched.
If there is an all data acquisition request from the data input / output unit 21, all the elements of all the chains are extracted with reference to the link information, and the result is returned to the data input / output unit 21.
[0042]
First, the processing procedure of the data search means 12 when a request for searching one piece of data is received will be described with reference to the flowchart shown in FIG.
A hash function is applied to the data to be searched to obtain a hash value (steps S21 and S22).
A bucket having the obtained hash value as an index is obtained (step S23).
The pointer to the element currently being processed is set as the pointer to the first element of the chain stored in the bucket (step S24).
[0043]
It is checked whether the pointer to the element currently being processed is zero (step S25).
If the pointer to the element currently being processed is zero (NO in step S25), the process ends, assuming that the data being searched has not been registered (step S26).
On the other hand, if the pointer to the element currently being processed is not zero (YES in step S25), the data stored in the element indicated by the pointer is extracted and if it is equal to the data to be searched (YES in step S27). Then, the process is terminated assuming that the data is the data being searched (step S28).
If the data of the element indicated by the pointer is not the data to be searched, the pointer to the next element stored in the element is replaced with the pointer to the element currently being processed (step S29), and the process returns to step S25. .
[0044]
Next, the processing procedure of the data search means 12 when receiving a request to retrieve all data will be described with reference to the flowchart shown in FIG.
Hereinafter, acquisition of all data using the rear link information will be described. However, acquisition of all data using the previous link information can be similarly performed. That is, in the procedure when the rear link information is used, the previous link information may be used instead of the rear link information.
[0045]
The index of the bucket currently being processed is initialized with the content of the post-link start index (step S31).
If the index of the bucket currently being processed is zero (YES in step S32), the processing ends as all buckets have been processed.
The data extracted so far is returned to the data input / output means 21 which has requested all data to be extracted.
[0046]
The pointer to the element currently being processed is initialized with the pointer to the head element of the chain list stored in the bucket (step S33).
It is checked whether the pointer to the element currently being processed is zero, and if it is an empty pointer (NO in step S34), there is no longer a pointer to the element currently being processed. The bucket currently being processed is replaced with the post-link information (step S35), and the process returns to step S32.
In this process, the extraction of the data relating to the chain list indicating the data having the same hash value (index) is completed, and the process proceeds to the processing of the data having the next hash value.
[0047]
On the other hand, if the pointer to the element currently being processed is not zero (YES in step S34), the data stored in the element indicated by the pointer is extracted (step S36).
The pointer to the next element stored in the element is obtained, set as the pointer to the element currently being processed (step S37), and the process returns to step S34.
[0048]
In response to a data deletion request from the data input / output unit 21, the data deletion unit 13 obtains a chain list in which the given data is registered, and the given deletion request data is registered in the chain list. If it is, it is deleted and the result is returned to the data input / output means 21.
The processing procedure of the data deletion unit 13 will be described with reference to the flowcharts shown in FIGS.
[0049]
A hash function is applied to the data to be deleted passed from the data input / output unit 21 to obtain a hash value (steps S41 and S42).
A bucket having the obtained hash value as an index is obtained (step S43). Hereinafter, the index obtained here is defined as i.
The pointer to the element currently being processed is set as the pointer to the first element of the chain stored in the bucket (step S44).
[0050]
It is checked whether the pointer to the element currently being processed is zero (step S45).
If the pointer to the element currently being processed is zero (NO in step S45), the process ends assuming that the data designated to be deleted has not been registered (step S46).
On the other hand, when the pointer to the element currently being processed is not zero (YES in step S45), unless the data stored in the element indicated by the pointer matches the data to be deleted (NO in step S47). Then, the pointer to the next element stored in the element is replaced with a pointer to the element currently being processed (step S48), and the process returns to step S45.
On the other hand, if the data stored in the element indicated by the pointer matches the data to be deleted (YES in step S47), since this data is the data to be deleted, the following processing is performed. .
[0051]
It is checked whether the element storing the data to be deleted is the head of the chain list. If the element is the head of the chain (YES in step S49), the pointer of the next element indicated by the element to be deleted is stored in the bucket, and this deletion target The element is deleted (Step S50).
For example, as shown in FIG. 11A, when deleting the first element (the element storing the value 1) in the chain list, the element following the element storing the value 1 (the element storing the value 2) ) Is stored in the bucket.
[0052]
If the element to be deleted is not the head of the chain list (NO in step S49), the pointer pointed to by the element immediately before the element storing the data to be deleted is set to the pointer to the element next to the element storing the data to be deleted. Then, the element to be deleted is deleted (step S51).
For example, as shown in FIG. 11B, when deleting the second element (the element storing the value 2) in the chain list, the element next to the element storing the value 2 (the value 3 is stored). The pointer to the element) is a pointer to the element next to the first element (the element storing the value 1).
[0053]
Next, after deleting the element to be deleted, if there is still an element in the chain list (NO in step S52), the deletion processing ends.
On the other hand, if there is no element in the chain list as a result of deleting the element to be deleted (YES in step S52), the following processing shown in the flowchart of FIG. 10 is performed.
[0054]
Assuming that the subsequent link information of the index i is j, the previous link information of the index j is replaced with the previous link information of the index i (step S53).
Assuming that the previous link information of the index i is k, the link information after the index k is replaced with the link information after the index i (step S54).
[0055]
If the content of the post-link start index is index i (YES in step S55), the post-link start index is replaced with post-link information of index i (step S56).
If the content of the previous link start index is index i (YES in step S57), the previous link start index is replaced with the previous link information of index i (step S58).
Finally, zero is stored in the subsequent link information and the previous link information of the index i (steps S59 and S60), and the deletion processing ends.
[0056]
For example, when the data stored at the head of the bucket with the hash value of 3 is deleted from the contents of the hash table storage unit 14 in the state shown in FIG. 4, the result becomes as shown in FIG.
[0057]
The present invention is not limited to only the above-described embodiment. Each function of the data management apparatus of the above-described embodiment is programmed and written in a recording medium such as a CD-ROM in advance, and this CD-ROM is stored in a medium drive such as a CD-ROM drive mounted on a computer. It is needless to say that the objects of the present invention can be achieved by installing these programs in a memory or a storage device of a computer and executing the programs.
In this case, the program itself read from the recording medium implements the functions of the above-described embodiment, and the program and the recording medium on which the program is recorded also constitute the present invention.
[0058]
In addition, as a recording medium for storing the program, a semiconductor medium (for example, ROM, nonvolatile memory card, etc.), an optical medium (for example, DVD, MO, MD, CD, etc.), a magnetic medium (for example, magnetic tape, flexible disk, etc.) ) May be used.
[0059]
Further, not only the functions of the above-described embodiment are realized by executing the loaded program, but also the above-described execution is performed by performing processing in cooperation with an operating system or another application program based on an instruction of the program. The case where the function of the form is realized is also included.
[0060]
When distributing to the market, the program is stored and distributed in a portable recording medium, or stored in a storage device of a server computer connected via a communication network such as the Internet, and another program is stored through the communication network. It can also be transferred to a computer. In this case, the storage device of the server computer is also included in the recording medium of the present invention. The functions of the above-described embodiments are implemented by installing a program on a portable recording medium or a transferred program into a recording medium built in a computer, and executing the installed program.
[0061]
【The invention's effect】
As described above, according to the present invention, when retrieving all data registered in the hash table, the data is registered regardless of the size of the hash table or the number of data stored in the hash table. A bucket can be searched at high speed.
[0062]
Furthermore, the bucket can be searched at high speed regardless of the size of the hash table or the number of data stored in the hash table, regardless of whether the data is retrieved in the normal or reverse order.
[Brief description of the drawings]
FIG. 1 is a block diagram of a computer system constituting an embodiment of a data management device according to the present invention.
FIG. 2 is a block diagram showing a functional configuration of a data management device according to the present invention.
FIG. 3 is an example of a data structure of a hash data storage unit.
FIG. 4 is a diagram for explaining data setting of hash data storage means.
FIG. 5 is a flowchart illustrating a processing procedure of a data registration unit.
FIG. 6 is an example in which new data is registered in the hash data storage means of FIG.
FIG. 7 is a flowchart illustrating a processing procedure of a data search unit when a search request for designated data is received.
FIG. 8 is a flowchart showing a processing procedure of a data search unit when a request to retrieve all data is received.
FIG. 9 is a flowchart illustrating a processing procedure of a data deleting unit.
FIG. 10 is a flowchart showing a processing procedure of a data deleting unit (continuation of FIG. 9).
FIG. 11 is an explanatory diagram when an element is deleted from a chain list.
FIG. 12 is an example when data is deleted from the hash table storage means in the state shown in FIG. 4;
FIG. 13 is a diagram for explaining a configuration of a conventional hash table.
[Explanation of symbols]
A: Previous link start index, P: Previous link information, B: Rear link start index, N: Rear link information, H: Hash table, E: Chain list element, Ev: Data value of element E or pointer to data , Ep ... Pointer to the next element in the chain list of element E, 10 ... server, 11 ... data registration means, 12 ... data search means, 13 ... data deletion means, 14 ... hash table storage means, 20 ... terminal, 21 data input / output means, 30 communication network.

Claims (7)

In a data management device that determines a storage location of data to be registered using a hash function, a hash table and information about data starting from a bucket on the hash table obtained by applying a hash function to data to be registered are stored. Hash table storage means including a chain list in which elements to be held are linked by a chain and link information for holding the association of the registered buckets on the hash table, and a bucket obtained by applying a hash function to the data to be registered And a data registering means for registering information on the data in the chain list shown in FIG. 2 and storing the information in the hash table storage means. If the data is registered in a new bucket on the hash table, Bucket to follow next bucket Data management apparatus is characterized in that so as to change the link information as. 2. The data management device according to claim 1, wherein the link information includes information indicating a position of a bucket to be followed before a bucket in which data is registered, and information of a bucket to be followed after a bucket in which data is registered. A data management device comprising: information indicating a position. 3. The data management device according to claim 1, wherein the link information is an index indicating a bucket. 4. The data management device according to claim 1, further comprising a data search unit for searching for registered data, wherein the data search unit stores data in the hash table from the link information when extracting all data. A data management device wherein a registered bucket is taken out and registered data is taken out from a chain list linked from the bucket. 5. The data management device according to claim 1, further comprising a data deletion unit for deleting registered data, wherein said data deletion unit deletes specified deletion data from said chain list. A data management device characterized in that when data registered in the chain list is exhausted, a bucket indicating the deleted data is deleted from the link information. A program for causing a computer to execute the functions of the data management device according to claim 1. A computer-readable recording medium on which the program according to claim 6 is recorded.

Images