JP2003030040A - Hush indexes of object database system and non-unique index management system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize high speed retrieval by enabling various kinds of index retrieval to perform to one table. SOLUTION: This object database system is provided with a storage device 100 in which a database is constructed, a storage device 101 to store hush pointers, a storage device 102 to store a group of hush buckets, an index ID generator 200 to generate index IDs, a retrieving device 201 to perform record retrieval in the database by using the hush pointers and the group of hush buckets and a storage operating device 202 to read/write the record to the database and to generate the hush buckets and to store them in the storage device 102 in the case of inserting the record into the database.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multi-hash index and non-unique index management system for an object database system that can speed up various index searches by adding a plurality of indexes to one table in an object database. .

[0002]

2. Description of the Related Art There are various databases on various operating systems, and each database needs a method that enables faster retrieval when retrieving and retrieving records inserted in the database. Looking at existing databases, most of the databases can be searched at high speed by index search.

[0003]

Conventionally, it is possible to add a plurality of or one index to one table, but since a non-unique index by hash search is not supported, various search keys can be used. There is a problem in that there are cases where it is necessary to perform a sequential search when performing a search that is used.

The present invention has been made to solve the above problems, and has a plurality of hash indexes and a non-unique index of an object database system capable of realizing high-speed search capable of performing various index searches on one table. The purpose is to provide a management method.

Another object of the present invention is to provide a multi-hash index and non-unique index management system for an object database system, which enables faster retrieval than sequential retrieval when retrieving a plurality of records.

[0006]

SUMMARY OF THE INVENTION A multiple hash index and non-unique index management system for an object database system according to the present invention is provided with an attribute of a table registered in the database and an index field of each record belonging to the attribute of the table. An index ID generating means for creating an index ID relating to a unique index or a non-unique index of each record based on the above, and the same index I having an address pointing to the record on the database.
A hash bucket storage unit that stores a hash bucket group connected to a hash bucket having D by a chain structure; a hash pointer storage unit that stores a hash pointer that defines a correspondence relationship between the index ID and the corresponding hash bucket; Retrieval means for high-speed retrieval using a hash pointer stored in the hash pointer storage means and a hash bucket stored in the hash bucket storage means for retrieval of records accumulated in the database, and inserting a record in the database At this time, storage operation means for generating the hash bucket and storing it in the hash bucket storage means is provided.

The multiple hash index and non-unique index management method of the object database system of the present invention is based on the attributes of the table registered in the database and the index field of each record belonging to the attribute of the table. A hash bucket group that creates an index ID related to a unique index or a non-unique index of a record, has an address that points to a record on the database, has the same index ID, and is connected by a chain structure, and the index ID and the corresponding By using the hash pointer that defines the correspondence relationship of the hash buckets, the records accumulated in the database can be searched, so that a unique table can be obtained. Or to achieve a non-unique index search and fast retrieval.

[0008]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below. The multiple hash index and non-unique index management system of this object database system is composed of a hash pointer and a hash bucket group as shown in FIG. The hash pointer is a correspondence table of index IDs and the corresponding hash buckets. The index ID is a positive value uniquely determined from the table name and the variable in which the index of the record is designated, and exists for each record by the number of indexes. The hash bucket group is a set of a plurality of hash buckets. The hash bucket has an address that points to a record on the database and is connected to a hash bucket having the same index ID by a chain structure.

Further, as shown in FIG. 2, when a record is inserted in the database, an index ID is generated from the index of the table to which the record belongs and the table name. When a plurality of indexes are added to the table, index IDs are generated by the number of indexes. Next, a hash bucket corresponding to the index ID is generated, and if the address value is set to NULL by referring to the address corresponding to the index ID of the hash pointer, the hash bucket generated at that address Is stored, and if the addresses of other hash buckets are already set, the hash buckets generated at the end of the chain of hash buckets are connected.

When searching for a record with an index, the index ID is generated, the address of the hash bucket is obtained from the hash pointer, and the chain of hash buckets is searched sequentially to find the corresponding record. In this way, a database configuration that allows index retrieval and non-unique index by attaching a plurality of indexes to one table becomes possible.

FIG. 3 is a block diagram showing each element and its structure for realizing the multiple hash index and non-unique index management system of the object database system of this embodiment, and a storage device 100 which is a database, Three storage devices, a storage device (hash pointer storage means) 101 that serves as a hash pointer and a storage device (hash bucket storage means) 102 that stores a hash bucket group, and an index ID generation device (index I) that generates an index ID.
D generation means) 200, a search device (search means) 201 for searching records in the storage device 100, and a storage operation device (storage operation means) 202 capable of reading / writing records in the storage device 100. ing.

The storage device 100 is an object database, in which a table having an attribute of "subscriber" exists, and the table of "subscriber" includes "telephone number", "surname", "first name" and "address". , And “age”. And there are three records that belong to this attribute (Record1, Record2, Record
rd3) is present. Two indexes for searching are attached to the "subscriber" table. Index A is a unique index that can be searched by subscriber number, and index B is a non-unique index that can be searched by age.

The index ID generation device 200 is a device that can generate an index ID from a table name and an index field of each record according to a certain rule. In the case of the example of FIG. 3, for example, a "subscriber" table is used. When creating an index ID related to the index A of Record 1 of the table 1, the table name “subscriber” and the telephone number field of Record 1 “047”
The index ID is created using "1-85-XXXXX". The search device 201 is a device that can perform a high-speed search for a record stored in the storage device 100 by using a hash pointer of the storage device 101 and a hash bucket group of the storage device 102.

The storage operation device 202 is a device that enables the insertion of records into the database of the storage device 100, the updating of records within the database, and the deletion of records. When inserting records into the database, Generates a hash bucket and stores it in the storage device 102.

Since the "X" in the storage device 101 is the index B, the index generation device 2 uses the value "31" of the age field of Record1 and record 2.
It is an index ID created by 00 and is "Y"
Is an index ID generated by the index ID generation device 200 from the age field “28” of Record 3 for index B, and “A” is the telephone number field “0” of Record 1 for index A.
471-85-XXXXX, the index ID created by the index ID generating apparatus 200, similarly "B" is created from the telephone number field "0471-85-XXXX" of record 2 for index A. The index ID, "C," is the telephone number field "0471-85- ○○ of record 3 for index A."
It is an index ID created from "O".

Next, the operation will be described. First, the insertion of a record into the storage device 100 will be described. Figure 4
FIG. 4 is a sequence diagram showing a processing procedure at the time of inserting a record into the storage device 100. Insertion of Record1 into the subscriber table of the storage device 100 is performed by the following procedure. That is, the storage operation device 202 acquires information about the subscriber table from the storage device 100 (step S1). As a result, the information about the subscriber table is acquired from the storage device 100. This information includes information that the index A and the index B are added to the subscriber table (step S2). The storage operation device 202 inserts a record (Record)
To create an index for index A against "0471 in the phone number field of the record.
"-85-XXXXX" and the table name "subscriber" are sent to the index ID generation apparatus 200 to acquire the index ID (step S3), which is received from the storage operation apparatus 202. The index ID is calculated from "0471-85-XXXXX" and "subscriber", and the result "A" is stored in the storage operation device 2
The data is sent to No. 02 (step S4).

The address of the hash bucket pointed to by the index ID = "A" is referenced from the hash pointer of the storage device 101 (step S5). If null is set, the hash bucket index ID = "A" points to the hash bucket. Set the address of, and if it is not null, attach the created hash bucket at the end of the chain of hash buckets. After setting the address of the hash bucket, the created hash bucket is inserted into the storage device 102 (step S6). Similar to step S3, the storage controller 202 creates an index for the index B for the record to be inserted, so that the age field of the record is "31" and the table name is "subscriber".
Is sent to the index ID generation device 200 to acquire the index ID (step S7). Similar to step S4, the index ID generation device 200 calculates an index ID from "31" and "subscriber" received from the storage operation device 202, and as a result sends "X" to the storage operation device 202 ( Step S8). From the hash pointer of the storage device 101, the index ID =
The address of the hash bucket pointed to by "X" is referred to (step S9), and if the null is set, the address of the hash bucket pointed to by the hash pointer index ID = "X" is set. Attach the hash bucket created at the end of the chain. After setting the address of the hash bucket, the storage device 1
The created hash bucket is inserted into 02 (step S10). After the index creation processing is completed, the storage operation device 202 inserts a record into the storage device 100 (step S11).

Next, a search for a record in the storage device 100 will be described. First, a description will be given of a record search using the index A (in the case of a unique index). In FIG.
The telephone number "047" of the subscriber table inserted in 0
1-85- ×× ○○ ”
Since the key of is the telephone number of the subscriber table,
A record is searched using the index A.
The search device 201 uses the index ID generation device 200 to create a table name “subscriber” and a telephone number “0471-”.
The index ID = “B” is acquired from “85-XXXXX”. After the search device 201 acquires the index ID, the index ID of the storage device 101 that is a hash pointer
= “B” is referred to obtain the address of the hash bucket. Then, the search device 201 displays the corresponding index ID.
Get the address to a record in the hash bucket of. This address is information indicating the address of the record in the storage device 100, and the record in the storage device 100 is referred to so that the value of the telephone number field is "0471-85.
-×× ○○ ”, it is determined that the record is a relevant record, the search is terminated, and when the value of the telephone number field is different, next in the hash bucket (the address indicating the next hash bucket in the chain of consecutive hash buckets ) Is obtained, the address to the record in the hash bucket is obtained, and the telephone number field is “0471-85-
XX ○○ ”is checked. By repeating this, the record search is completed.

Next, a search for a record using the index B (in the case of a non-unique index) will be described. In FIG. 3, when the search device 201, which is a search device, searches for the age “31” in the subscriber table inserted in the storage device 100, the index B key is the age of the subscriber table. A record is searched using B. The search device 201 uses the index ID generation device 200 to calculate the index ID = from the table name “subscriber” and age “31”.
Get "X". Search device 201 is the index ID
After the acquisition, the address of the hash bucket is acquired by referring to the index ID = “X” of the storage device 101, which is the hash pointer. Then, the search device 201 acquires the address to the record in the hash bucket of the corresponding index ID. This address is information indicating the address of the record in the storage device 100. If the value of the age field is "31" by referring to the record in the storage device 100, it is determined that the record is a corresponding record. When the values of the age fields are different, next in the hash bucket is acquired and records are searched sequentially. By searching this chain of hash buckets, a record in which the value of the age field of the subscriber table is "31" (see FIG. 3).
It is possible to search all Record1 and Record2) in.

As described above, according to this embodiment,
Since it is possible to add multiple indexes to one table, it is possible to perform high-speed search using the index of the record by various approaches according to the number of indexes added to one table, and to one table. This has the effect of performing various index searches.

In addition, a non-unique index is supported, and in the example of FIG.
Assuming that 0 records are inserted and 100 of them have an age value of 31, there are 100 to 10,000 record searches required for sequential search, but 100 records for index search. It is sufficient to search about the number of records, and when searching a plurality of records, there is an effect that a search can be performed faster than a sequential search.

In the embodiment described above, the number of tables registered in the database is one for the sake of convenience, but it is also applicable to the case where a plurality of tables are registered in the database. Also, an example in which the number of keys used in the index is one is shown, but the present invention is also applied to the case where there are a plurality of keys for one index.

[0023]

As described above, according to the present invention, an index ID relating to a unique index or a non-unique index of each record is created, and an address pointing to a record on the database is provided.
So that records stored in the database can be searched for by using a group of hash buckets connected by the same chain structure and a hash pointer that defines the correspondence relationship between the index ID and the corresponding hash bucket. Since it is configured, there is an effect that a unique or non-unique index search and high-speed search can be performed on one table.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a structure of respective records on a hash pointer, a hash bucket, and a database and their relations according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a process between a hash pointer and a hash bucket at the time of inserting a record according to the embodiment of the present invention.

FIG. 3 is a block diagram showing each element and its structure for realizing the multiple hash index and non-unique index management method of the object database system according to the embodiment of the present invention.

FIG. 4 is a sequence diagram showing a processing procedure at the time of inserting a record into the storage device according to the embodiment of the present invention.

[Explanation of symbols]

100 ... Storage device (database), 101 ... Storage device (hash pointer storage means), 102 ... Storage device (hash bucket storage means), 200 ... Index ID generation device (index ID generation means), 20
1 ... Search device (search means), 202 ... Storage operation device (storage operation means).

Claims (4)

[Claims] 1. An index ID for creating an index ID relating to a unique index or a non-unique index of each record based on the attribute of the table registered in the database and the index field of each record belonging to the attribute of the table. Generating means and an address pointing to a record on the database,
Hash bucket storage means for storing a hash bucket group connected to a hash bucket having the same index ID by a chain structure, and hash pointer storage for storing a hash pointer defining a correspondence relationship between the index ID and the corresponding hash bucket. Means, a search means for performing a high-speed search for a record stored in the database using the hash pointer stored by the hash pointer storage means and the hash bucket stored by the hash bucket storage means, and in the database Storage operation means for generating the hash bucket when inserting a record and storing the hash bucket in the hash bucket storage means; And non-unique index management method. 2. An index for search including a non-unique index is added to the table,
The index ID generation means relates to a unique index or a non-unique index based on an index for searching including the non-unique index, an attribute of the table, and an index field of each record belonging to the attribute of the table. The multiple hash index and non-unique index management method of the object database system according to claim 1, wherein an index ID is created. 3. The search means uses the hash pointer stored by the hash pointer storage means and the hash bucket stored by the hash bucket storage means based on the index ID generated by the index ID generation means, and the index 3. The multi-hash index and non-unique index management method of the object database system according to claim 2, wherein a match is found in the ID index field and the corresponding record is searched. 4. The storage operation means is the index I of the record generated by the index ID generation means.
Based on D, the hash pointer stored in the hash pointer storage means is used to generate the index ID.
Refers to the address of the hash bucket pointed to, if the null is set, set the address of the hash bucket pointed to by the index ID, if not null, insert the hash bucket created at the end of the chain of the hash bucket, The multiple hash index and non-unique index management method of the object database system according to claim 2, wherein the record is inserted into the database.