JP2005215869A - Data storage and retrieval system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a system capable of facilitating the management of an overflow block and realizing a high speed access. <P>SOLUTION: This data storage and retrieval system for storing data in a computer and retrieving the data is provided with: data records having data items including main key items; a primary block 11 and an overflow block 12 in which the data records are stored in the order of main keys, location table records for storing the address of the primary block 11; a location table 10 in which the location table records are stored in continuous regions in the order of the main keys; a final pointer showing the last of the use region of the location table 10; an overflow block management table 20 for managing the address of the overflow block; and an overflow block final pointer showing the last of the use region of the overflow block management table 20. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

The present invention relates to a system for storing and retrieving data using a computer.

  The conventional database storage and retrieval method by computer is "Jeffrey D. Ullman, Kunii et al.," The Principle of Database System ", 1st edition, Japan Computer Association, May 25, 1985, p45-71. , “Samuel Leffler et al./Akira Nakamura et al.,“ Design and Implementation of UNIX4.3BSD ”, Maruzen Co., Ltd., June 30, 1991, p193-191” and “Michael J. Folk et al./Hiroyuki Enomoto, As described in "bit separate volume file structure", Kyoritsu Publishing Co., Ltd., June 5, 1997, p169-191 ", it basically used a hierarchical index.

  The present inventor introduces the concept of a location table and an alternative key table in place of the conventional hierarchical index, simplifies complicated processing associated with index processing, and searches the location table itself by binary search, etc. By using this technique, a “data storage and retrieval method” has been invented (see Japanese Patent No. 3345628 and US Pat. No. 6,654,868), which can ensure high speed and easy maintenance. In this “data storage retrieval method”, an overflow block is connected to a primary block, and further, an overflow block is connected to an overflow block.

  Furthermore, the present inventor invented the “database reorganization system and database” (hereinafter referred to as “database reorganization system”, PCT / JP03 / 11592). We have invented a mechanism that allows reorganization of a database invented in the “data retrieval and storage system” while the database is in operation. Furthermore, it has been invented that an efficient reorganization is possible by adding an alternative key location table to an alternative key table. Further, as an existing invention related to the present invention, a “database accelerator function (PCT / JP03 / 13443)” was invented.

The inventor has also invented a “database accelerator function (hereinafter referred to as“ accelerator function ”; PCT / JP03 / 13443)”. This allows the accelerator system to have a functional location identical to the primary system's location table and alternate key location table, the Fland location table and the Fland alternate key location table. As a result, the database can be accessed in parallel.

<1> The present invention is to efficiently manage a storage area.
<2> Another object of the present invention is to measure the performance improvement of the database system.

The present invention relates to a data storage and retrieval system for storing and retrieving data in a computer, a data record having a data item including a primary key item, a primary block and an overflow block for storing the data record in the order of the primary key, A location table record that stores the address of the primary block, a location table that stores the location table record in a continuous area in the order of the primary key, a final pointer that indicates the end of the used area of the location table, and overflow A data storage / retrieval system comprising an overflow block management table for managing block addresses and an overflow block last pointer indicating the end of the area used by the overflow block management table A.

<1> Basic configuration of data storage and retrieval system and solution of problem The data storage and retrieval system stores and retrieves data in a computer. For example, as shown in FIG. A data record, a primary block 11 and an overflow block 12 that store the data records in the order of the primary key, a location table record that stores the address of the primary block 11, and a location table record in the order of the primary key The location table 10 stored in the continuous area, the final pointer FP indicating the end of the used area of the location table 10, the overflow block management table 20 for managing the address of the overflow block 12, and the overflow block management table And an overflow block last pointer OF indicating the end of the used area of the bull 20.

  The overflow block management table record (or overflow block management table entry) includes the address of the overflow block 12 managed by the record and the minimum and maximum values of the primary key value of the overflow block 12. It consists of the address of the next overflow block management table record. Further, a pointer to the overflow block management table record is provided in the location table 10. The overflow block management table record stores the address each time an overflow block 12 occurs. If there are two or more overflow blocks 12 for one primary block 11, the overflow block management table record is the first one. Chain the overflow block management table record to the overflow block management table record of the second overflow block, the second to the third, and so on. In order to indicate the end of the record of the overflow block management table 20 being used, an overflow block management table pointer 20 is provided. The above description is the same for the substitute key overflow block.

  The data storage and retrieval system can solve the following problems. First, the overflow block 12 is connected to the primary block 11, and the overflow block 12 is further connected to the overflow block 12. The method of connection is the immediately preceding primary block 11 or overflow block. If 12 holds the address of the subsequent overflow block 12, it must search for the overflow block 12 one after another to find the desired record. In that case, the location table 10 and the alternative key location table are arranged in a high-speed random access memory, and the block and the alternative key block are relatively inexpensive such as a hard disk, but the access speed is relatively slow. When storing in the medium, if the number of overflow blocks 12 or alternative key overflow blocks increases, the overflow blocks 12 must be searched one after another in the slow storage medium, and the target overflow block 12 is stored. There is a problem that it takes a long time to search. For example, if a high-speed random memory can be accessed in 100 ns, while an overflow block requires an access speed of 10 ms, the difference is 100,000 times. This problem becomes more problematic as the number of connected overflow blocks 12 increases.

  When the overflow block management table 20 is used, it is possible to identify which primary block 11 or overflow block 12 stores the target record simply by searching the overflow block management table 20. Access to block 11 or overflow block 12 is only required once.

  Next, when there are a plurality of overflow blocks 12 and a record is inserted into a relatively forward overflow block 12 or primary block 11, all the records behind that record are moved. This must increase the load required. Regarding this problem, the movement of the record may be minimized by inserting a new overflow block 12 between the primary block 11 and the overflow block 12 or between the overflow block 12 and the overflow block 12. It becomes possible.

<2> Data record A data record always has one unique primary key and zero or one or more non-unique keys (which may be unique as a result) (substitute keys). In the case of an employee database, the primary key is a code that can identify the employee, such as an employee code, and the alternative key is the name, date of birth, etc. I do not care. In addition, for a record in which there is no item that has a meaningful primary key, a sequential number in the storage order may be assigned and used as the primary key.

<3> Primary Block The primary block 11 is a block managed by the location table entry, and there is only one primary block 11 for each location table entry. These blocks store data records in the order of primary keys.

<4> Location Table The location table LC10 reserves location table records (or location table entries) in a continuous area in advance. For example, as shown in FIG. 2, the location table record includes an overflow block management table that manages an overflow block 12 connected to the primary block 11 when there is an address of the primary block 11 and an overflow block. -It has an entry address. Here, the continuous area is a logical order, and may be separated in a physical area. When the physical area is separated, it can be treated as logically continuous using the address conversion table (the same applies hereinafter). The last pointer FP (101 in FIG. 5) is used to indicate the end of the used area of the location table. The last pointer FP can be used to know the last address of the area where the location table record is stored.

  When a data record cannot be added to the last primary block 11, a new primary block 11 is added after that and the data record is stored. The address location table record of the added primary block 11 is written in the location table LC10, and the position of the final pointer FP is moved backward by one.

<5> Overflow block When a data record is added to the primary block 11, if the primary block 11 is full, the overflow block 12 is connected to the primary block 11 to store the data record. To do. It is possible to further connect the overflow block 12 to the overflow block 12. Here, concatenation does not mean that they are physically connected, but means that each is pointed by an address pointer. In the “data storage search method”, the primary block 11 holds the address of the first overflow block 12, and the first overflow block 12 holds the address of the second overflow block 12. Further, in this specification, a new method of connection using the overflow block management table 20 is described.

<6> Access by primary key A method for searching for a record having a primary key is to perform a binary search between the first record of the location table 10 and the location table record pointed to by the final pointer. The target primary block 11 is found, the inside of the primary block 11 is searched, and the target record is found. When the overflow block 12 is connected to the primary block 11, the overflow block 12 is also a search target.

<7> Alternative Key Table The alternative key table (20A, 20B, 20C) includes an alternative key location table (AALC, ABLC, ACLC) and an alternative key block (22A, 22B) as shown in FIG. 22C). The alternate key location table and alternate key block are similar to the location table and primary block for the primary key. The substitute key block stores substitute key records (or substitute key entries) composed of substitute key values and primary key values in the order of substitute key values. The alternative key location table has an alternative key location table record (or an alternative key location table entry) in which the address of each alternative key block is described in a continuous area. The substitute key location table is secured in a continuous area in advance. The substitute key final pointer FP (29A, 29B, 29C) is used to indicate the end of the used area of the substitute key location table. When adding an alternate key record, if an alternate key record with an alternate key value greater than the alternate key value of an existing alternate key record is stored in the last alternate key block and cannot be stored in that alternate key block A new substitute key block is created, and the substitute key record is stored in the substitute key block.

<8> Access by alternative key The method of searching for a record having a certain alternative key is to search for the first record of the alternative key location table and the alternative key location table record pointed to by the alternative key final pointer. Binary search between them, find the desired alternate key block, and search within the alternate key block to find the alternate key record with the desired alternate key. If an alternative key overflow block is concatenated to the alternative key block, the alternative key overflow block is also searched.

  Next, the location table LC is binary searched by the primary key of the alternative key record, the target primary block is found, and the target record is found from the primary block and the overflow block concatenated therewith. In this example, the location table is binary searched. However, in the “data storage search method”, the alternative key entry can have the address or number of the primary block. Thus, if the alternate key entry has the address or number of the primary block, it is possible to access the block directly without performing a binary search of the location table.

  Since the substitute key is non-unique, there may be a plurality of data records having the same substitute key value. In this case, if the next substitute key record in the substitute key block has the same substitute key value, the same operation as described above is repeated. When a plurality of alternative keys exist, the same number of alternative key tables as the types of the alternative keys are created and used.

<9> Storage Using Overflow Block Management Table The overflow block management table 20 has an overflow block management table record (or overflow block management table entry) and manages the overflow block 12. It is. For example, as shown in FIG. 3, the overflow block management table record is one of the address of the overflow block 12 managed by the entry and the minimum value and the maximum value of the primary key value of the overflow block 12. Or both and the address of the management table record of the next overflow block 12. As shown in FIG. 2, the location table 10 has an item for storing the address of the overflow block management table record. In the location table 10, if the overflow block 12 exists in the item for storing the address of the overflow block management table record, the address of the record is stored, and there is no overflow block. Stores a specific value that can be identified as non-existent to indicate that there is no overflow block. The “next overflow block management table record address” of the overflow block management table record can identify that the next overflow block does not exist when the next overflow block 12 is not concatenated. Indicates that a specific value is stored and does not exist.

  As with the location table 10, the overflow block management table 20 is secured in a continuous area in advance. Here, the continuous area is a logical order, and may be separated in a physical area. In such a case, it can be treated as logically continuous using the address conversion table. The same applies to the following description. In this way, using the address translation table, it is possible to use physically non-contiguous areas, so it is possible to secure incremental areas when the area initially reserved is insufficient. Yes, it can be avoided that the system becomes inoperable due to insufficient area. That is, it is also possible to take a method in which areas are secured each time an overflow block 12 occurs and those areas are pointed to one after another. It is preferable to describe in the database system that the overflow block management table 20 is used.

  In order to identify where the overflow block management table record is used, an overflow block management table pointer (OFP) is provided to point to the next address of the last overflow block management table record. To do. This will be described with reference to FIG. OFP indicates the beginning of the overflow block management table 20 at the beginning. For the numbers in the overflow block management table record, the first number is the number of the overflow block management table record, and the number in parentheses is the number of the overflow block managed by the record.

  When an overflow block occurs for block 1 of primary block 11, the first record 0 of the overflow block management table is reserved for overflow block 1-1. The address of overflow block management table record 0 is stored in location table record 1. Next, the address of the overflow block 1-1 is stored in the overflow block management table record 0. In FIG. 1, the above address storage is indicated by a dotted line. Solid lines indicate that the location table 10 points to the primary block 11 and the primary block 11 to the overflow block 12 as before. The link indicated by the solid line, that is, the primary block 1 holding the address of the overflow block 1-1 may or may not be provided. Next, the maximum value and / or minimum value of the record stored in the block 1-1 of the overflow block 12 is stored in the overflow block management table record 0. The OFP is changed to indicate the head of the overflow block management table record 1.

  Since the next overflow block does not exist in the overflow block 1-1, a specific value is entered in the "address of the management table record of the next overflow block" of the overflow block management table record 0. , So that the next overflow block can be identified.

  In FIG. 1, the overflow block 4-1 occurs next. The operation accompanying this generation is the same as described above. When the overflow block 4-2 occurs next, the overflow block 4-2 follows the overflow block 4-1, so that the "next overflow block management table" in the overflow block management table record 1 is displayed. “Address of record” is the address of overflow block management table • record 2 and overflow • Block management table • Record 2 “address of management table record of next overflow block” is a specific value . Other operations are the same as described above. FIG. 1 shows a point in time when the addition up to the overflow block 4-3 is performed.

<10> Data Record Access Using Overflow Block Management Table Next, access using the overflow block management table 20 will be described with reference to FIGS. FIG. 2 is an example of a location table record. FIG. 3 is an example of an overflow block management table record. When the overflow block management table is not used, the target primary block 11 is found by performing a binary search on the location table 10, and the target data is detected from the primary block and the overflow block 12 connected thereto. Find a record.

  A binary search of the location table reads an address between the first address of the location table and the address pointed to by the last pointer, and the primary key is in the location table record for that address, or the higher address, or the lower address Check if it is at the address. If it is at the upper address or the lower address, the intermediate address is further searched by a binary search to find the primary block 11 containing the primary key and the overflow block 12 connected thereto.

  Access when the overflow block management table 20 is used is as follows. The location table is binary searched with the primary key value (target key value) of the target data record to find the primary block 11 storing the target record and the overflow block 12 concatenated therewith. The location table entry contains the minimum and / or maximum primary key value of records stored in the primary block managed by the entry and the entire overflow block connected to the primary block. Is stored, it is possible to find the target primary block and the overflow block concatenated with it by comparing the value with the target key value. For example, it is assumed that the primary block 3 in FIG. 1 is the target primary block. Next, the symbol indicating the presence or absence of the overflow block described in the location table record 3 is examined. In this case, since there is no overflow block, the primary block 3 is accessed to find the target data record.

  Next, a case where the target primary block is block 4 will be described. In this case, looking at the location table record 4, it can be identified that an overflow block exists. The overflow block management table record is an overflow block management table record 1. Therefore, the overflow block management table record 1 (overflow block 4-1) is referred to. As shown in FIG. 3, the record stores both the minimum value and the maximum value of the primary key value of the record stored in the overflow block 4-1, or one of them. The minimum value and the maximum value are compared with the target primary key value to determine whether or not a data record is included in the overflow block. If the target primary key value is smaller than the minimum value of the overflow block 4-1, the target record is present in the primary block. If the target primary key value is equal to or between the minimum and maximum values of the overflow block management record 1, the target record is stored in the overflow block 4-1. If only one of the minimum or maximum primary key values is stored in the overflow block management table record, the minimum or maximum logical block before or after the overflow block management table record By referring to the value on the overflow block management table 20, it is possible to determine whether or not the target record is stored in the overflow block.

  When the target primary key value is larger than the maximum value of the overflow block 4-1, the overflow block management table record 2 (overflow block 4-2) is referred to. Again, if there is no target data record, the overflow block management table record 4 (overflow block 4-3) is referred to. If it is not included in any of them, there is no data record. If the primary key value is contained in the overflow block 4-3, the overflow block 4-3 is accessed and the target data record is found therein. Even if the record is included in the primary block or overflow block, the record may not exist due to an error in the target key value. Thus, the access to the overflow block is only once regardless of the number of overflow blocks. When the overflow block is stored in a low-speed device, high-speed access can be realized.

  In the above description, the case of search is described, but updating, adding, and deleting of data records can be performed in exactly the same manner. The data record is updated by updating the data record found by the search. In the deletion, the data record found by the search is deleted, and the data record behind the deletion record is moved forward by the area occupied by the deletion record. For the addition, a location for storing the data record to be added is found by using a search method, and after the data record after the storage location is moved backward, the data record is stored (added). Update, add, and delete are executed exclusively for the primary block and the overflow block connected to it.

  The above description has been given with respect to a method for efficiently managing and accessing the overflow block using the location table 10 and the overflow block management table 20, but the alternative key location table and the alternative key overflow table are described. The same effect can be produced by using the alternative key overflow block management table for the block management table. As shown in FIG. 10, an overflow block management table (20A, 20B, 20c) is provided for each alternative key location table (ALA0, ALB0, ALC0). Further, an overflow block management table pointer OFP (20A1, 20B1, 20C1) is provided for each overflow block management table. These usage and access methods are the same as those in the case of the location table and block, and will be omitted because they overlap.

<11> Inserting an overflow block in the middle If there are many overflow blocks and a data record is inserted into a relatively forward overflow block or primary block, all records after that record are moved backward. However, by inserting an overflow block in the middle, there is no need to move the data record after that data record and insert the data record. can do. The operation when an overflow block is inserted in the middle will be described with reference to FIG. A new overflow block 4-0 is secured. The primary block 4 points to the overflow block 4-1, but this is changed so that the primary block 4 points to the overflow block 4-0 and the overflow block 4-0 identifies the overflow block 4-1. Point to it. Next, among the records stored in either the primary block 4 or the overflow block 4-1, or both, necessary ones are moved to the overflow block 4-0. If it is simply done, it is sufficient to move the record to make room for storing the inserted record, but it is possible to move so that the storage ratio of the primary block and the overflow block before and after are the same. In that case, the record and the location table record are also rewritten. An overflow block management table record 5 is added. The overflow block management table record 5 manages the overflow block 4-0. The location table record 4 points to the overflow block management table record 1 (the management target is the overflow block 4-1). This is changed so that the location table record 4 points to the overflow block management table record 5 (overflow block 4-0). Write an address so that overflow block management table record 5 points to overflow block management table record 1 (overflow block 4-1). There is no change from 4-1 to 4-2 and from 4-2 to 4-3. Further, the overflow block management table record 5 (overflow block 4-0) stores the minimum value and / or maximum value of the primary key value of the record stored in 4-0.

Although the above description has been given with respect to a method for efficiently inserting the overflow block 12 using the location table 10 and the overflow block management table 20, the alternative key location table and the alternative key overflow block are described. A similar method can also be used when an alternative key record is inserted using the management table. This means that an alternate key overflow block is added when an alternate key record is inserted, between the alternate key block and the alternate key overflow block, or between the alternate key overflow block and the alternate key overflow block. By doing so, it is possible to efficiently insert an alternative key record.

  Below, the example of application to the reorganization system of a data storage search system is demonstrated.

<1> Database Reorganization System The “database reorganization system” will be described with reference to FIG. The reorganization is to eliminate the overflow block, secure an appropriate initial storage rate, and eliminate fragmentation. The elimination of the overflow block is as follows. When a large number of overflow blocks 12 and 13 are connected to the primary block 11, when attempting to insert a record into these blocks, a data record straddling the primary block 11 and the overflow blocks 12 and 13 Must be arranged in the order of primary key values, which increases the number of data records to be moved. Even when a data record is searched, the search must be performed across a plurality of blocks, resulting in a decrease in efficiency. In order to avoid this, the overflow blocks 12 and 13 are eliminated and the primary block 11 is made.

  Ensuring an appropriate initial storage rate is as follows. If an appropriate free space is provided in the primary block in advance, even if a data record is inserted, the data record can be inserted without immediately adding the overflow block 12. However, if data record insertion is repeated many times, it becomes a free space that deviates from the appropriate initial storage rate. This is returned to the initial state. Defragmentation is similar to ensuring an appropriate initial storage ratio, but primary blocks 11 and overflow blocks 12 that are no longer needed are deleted, or primary blocks or overflow blocks that have a low storage ratio are integrated. For example, the primary block and the overflow block are uniformly used. Although the primary block 11 and the overflow block 12 have been described here, the same applies to the alternative key block and the alternative key overflow block.

  In the reorganization of the location table, the primary block, and the overflow block, two location tables 10 are prepared, that is, a working location table LC and a new location table LN. Further, for each location table, one reorganization pointer RPLC is provided for the working location table to indicate how far the reorganization has been completed. One reorganization pointer RPLN is provided for the new location table. Further, the reorganization pointer RPLN for the new location table can be replaced by the final pointer FP.

  FIG. 7 illustrates the elimination of the overflow block. The block pointed to by the working location table LC includes a primary block 11 and overflow blocks 12 and 13. Since the first block of the working location table LC consists only of the primary block 0, it is transferred to the first location table record 0 of the new location table LN. Next, when the primary block 1 is seen, the overflow blocks 1-2 and 1-3 are connected. The primary block 1 is transferred to the location table record 1 of the new location table LN. Next, the connection of the overflow block 1-2 is cut off, and the address of the overflow block 1-2 is written in the record 2 of the new location table LN to make it the primary block. Similarly, the overflow block 1-3 also writes the address of the overflow block 1-3 to the record 3 of the new location table LN to make it the primary block.

  Similarly, FIG. 7 shows a point in time when overflow blocks are separated one after another and the overflow block resolution up to the primary block 3 managed by the record 3 of the working location table LC is completed. The reorganization pointer RPLC of the working location table points to the next address of the record 3 of the working location table LC. The relocation pointer RPLN of the new location table points to the next address of the record 6 of the new location table.

  Next, to secure an appropriate initial storage ratio and to eliminate fragmentation, target multiple blocks at one time. Between those primary blocks and overflow blocks, between blocks that do not have an appropriate initial storage ratio. The data record is moved so that the data record is properly stored in the block. Depending on the situation, the block is deleted or added. Here, the primary block and the overflow block have been described, but the same applies to the alternative key block and the alternative key overflow block.

  It is possible to access the database during reorganization. First, search is described. The search determines whether the primary key value of the record stored in the block pointed to by the record pointed to by the reorganization pointer RPLC is larger or smaller than the target primary key value. If it is smaller, a new location table LN is used to search for the target record by performing a binary search between the head and the reorganization pointer RPLN. If it is larger or equal, the target record is searched by performing a binary search using the current location table LC while the reorganization pointer RPLC and the final pointer FP are pointing. Although the case of search has been described here, update, addition, and deletion can be performed in the same manner. Reorganization and access in the alternate key table can be performed in almost the same manner as in the location table and block.

<2> Application Method of Database Reorganization System A method for eliminating the record in the overflow block management table when reorganization is performed by the “database reorganization system” will be described. The “database reorganization system” describes a method of simultaneously performing three methods of eliminating an overflow block, eliminating fragmentation, and ensuring an appropriate initial storage rate. Among them, since the overflow block management table 20 is added in the present invention when the overflow block is eliminated, a method of performing reorganization in this combination will be described with reference to FIG. In this description, the elimination of the overflow block will be described. However, when the reorganization is actually executed, it is preferable to simultaneously eliminate the fragmentation and ensure an appropriate initial storage ratio.

  To perform the reorganization, a new location table LN is provided for the working location table LC, and the location table record is transferred from the working location table to the new location table. At this time, if an overflow block exists, the overflow block is made a primary block, and a new location table record is created in the new location table. Accordingly, the number of records in the new location table is increased from the number of records in the working location table. FIG. 8 shows a state in which the reorganization is performed so that the overflow block becomes the primary block, and the reorganization is completed up to the record 4 of the working location table LC and the record 9 of the new location table LN. FIG.

  Reorganization starts at working location table record 0. It can be determined by referring to the working location table record 1 that the location table record 1 is connected to the overflow block 1-1. Since the location table record 1 points to the overflow block management table record 0, the address of the overflow block 1-1 pointed to by the overflow block management table record 0 is assigned to the new location table LN. Store in record 2. Further, the overflow block management table record 0 is not necessary.

  Next, the working location table record 2 is transferred to the new location table record 3. Similarly, the working location table records 3 and 4 are transferred to the new location table records 4 and 5. Four overflow blocks are connected to the primary block 4 managed by the working location table record. In the overflow block management table, the overflow block management table records are stored in the order of 5, 1, 2, and 4. As described above, the address of the overflow block pointed to by the overflow block management table record is stored in the new location table records 6, 7, 8, and 9. As a result, the overflow block management table records 5, 1, 2, and 4 become unnecessary. As described above, FIG. 8 shows a point in time when the reorganization is completed so far.

  Here, in addition to the new location table, a new overflow block management table is provided for the working overflow block management table. The records (0, 1, 2, 4, 5) in the working overflow block management table in FIG. 8 are records that are no longer necessary because the overflow block has been eliminated by reorganization. As described above, the overflow block management table record is disseminated with the reorganization. On the other hand, before the reorganization is completed, an overflow block may further occur for the overflow block that has already been reorganized.

  In order to manage these newly generated overflow blocks, a new overflow block management table is provided. The new overflow block management table stores the address, maximum value, and minimum value of the overflow block when an overflow block occurs for the primary block managed by the new location table. If an overflow block occurs for the primary block managed by the working location table, the working overflow block management table is used. When the reorganization is completed, the working location table and the working overflow block management table become unnecessary.

In the above, reorganization of location table and block and overflow block was explained. By using this method, alternative key location table and alternative key block, alternative key overflow block, alternative key -Overflow block management table can be reorganized in exactly the same way.

  An example of application of the data storage / retrieval system to an accelerator will be described below.

<1> Accelerator System An accelerator system will be described with reference to FIG. The principle of the accelerator system is as follows. Perform a binary search on the location table or alternate key location table to find the desired record. When performing a binary search on the location table, you will find the bisection point many times, which is more than the number of times you search for a record in the primary block or overflow block. Is common. Also, the possibility of requesting records in the same primary block or overflow block from multiple processes at the same time is quite low. Therefore, if a plurality of copies of the location table and the alternative key location table are held and a binary search can be performed on each of them in parallel, a large number of access requests for records can be executed.

  FIG. 10 shows a case where there is one accelerator system. Accelerator system 3 has a location table (Fland location table) and an alternate key location table (Fland alternate key location table), but the primary block, overflow block, and alternate block Does not have a key block or alternate key overflow block. The accelerator system's Fland location table is functionally equivalent to the primary system 1 location table. Similarly, the alternate key location tables ALA1, ALB1, and ALC1 of the accelerator system 3 are functionally equivalent to the alternate key location tables ALA0, ALB0, and ALC0 of the primary system 1. Each record of the rand location table 16 of the accelerator system 3 points to the same block as the record of each location table 10 of the primary system.

  In the accelerator system, when there is an access request, in the case of the primary key, a binary search is performed on the land location table to find the target primary block and the overflow block concatenated with it, and within that block. Ask the primary system to retrieve the records for For alternate keys, perform a binary search of the alternate alternate key location table to find the desired alternate key block and its associated alternate key overflow block, and the alternate held by the primary system Find the desired alternate key record from the key block, and use that alternate key record to perform a binary search of the Fland location table to find the desired record. Although the search method has been described here, it is possible to update, add, and delete records by applying this method. Also, in the case of an alternate key, binary search of the land location table is performed based on the alternate key record, but the alternate key record holds the address and block number of the primary block. It is not necessary in some cases. In this way, it is possible to increase the amount of processing by searching and updating records in parallel with a plurality of accelerator systems.

  In FIG. 10, the accelerator system has one location table and three alternative key location tables, which is the same number as the primary system. This is called a symmetric system in "Accelerator". On the other hand, for example, an accelerator system that holds only two types of location table and alternate key location table is assumed, and only the location table, only the alternate key location table, It is possible to create an accelerator system such as This is called an asymmetric system. With respect to the accelerator system, the primary block and the overflow block, the alternative key block and the alternative key overflow block can be handled in substantially the same manner.

<2> Application to Accelerator System Next, application to a system using the “accelerator function” will be described with reference to FIG. The primary system 1 has an overflow block management table 20 for the location table 10 and an alternative key overflow block management table 20A for the alternative key location table ALA0. The key location table ALB0 has an alternative key overflow block management table 20B, and the alternative key location table ALC0 has an alternative key overflow block management table 20C. In addition, the overflow block management table is provided with an overflow block management table pointer 201, and an alternative key overflow block management table pointer 20A1 is provided for the alternative key overflow block management table 20A. Alternate key overflow block management table pointers 20B1 and 20C1 are provided. Furthermore, it has final pointers (101, 10A1, 10B1, 10C1).

  In the accelerator system 3, a Fland overflow block management table 21 is provided for the Fland location table 16, and a Fland substitute key is provided for each of the ALA1, ALB1, and ALC1 Fland substitute key location tables. Provide overflow block management tables 21A, 21B, and 21C. For each of the alternative key / overflow / block management tables 21A, 21B, and 21C, the alternative key / overflow / block management table / pointer 21A1, 21B1, 21C1 is provided. Furthermore, it has a final pointer (161, 16A1, 16B1, 16C1).

  In the accelerator system 3, when a change occurs in the location table or the alternate key location table in the primary system 1, the change is notified to the accelerator system 3, and the accelerator system is notified. In step 3, the relevant location table or the alternative key location table is changed. By revising this as follows, the accelerator system 3 can be operated using the overflow block management table and the alternative key overflow block management table of the present invention. In primary system 1, location table 10, overflow block management table 20, final pointer 101, overflow block management table pointer 201, alternative key location table (ALA0, ALB0, ALC0), alternative key location Change to one of the last pointer (10A1, 10B1, 10C1) of the table, the alternative key overflow block management table (20A, 20B, 20C), or the alternative key overflow block management table pointer (20A1, 20B1, 20C1) When this occurs, the changed part is notified to the accelerator system 3. In the accelerator system 3, based on the notification, the Fland location table 16, the Frand overflow block management table 21, the Fland final pointer 161, the Frand overflow block management table pointer 211, the Fland alternative key, Location table (ALA1, ALB1, ALC1), Fland alternative key location table last pointer (16A1, 16B1, 16C1), Fland alternative key overflow block management table (21A, 21B, 21C), Fland alternative key The change portion is changed for any of the overflow block management table pointers (21A1, 21B1, 21C1).

  In this way, the change part is notified from the primary system 1 to the accelerator system 3, and the change is immediately applied by the accelerator system 3, so that the accelerator location table 16 of the accelerator system 3, Fland overflow block management table 21, Fland last pointer 161, Fland overflow block management table pointer 211, Fland alternative key location table (21A, 21B, 21C), Fland alternative key location table final pointer (21A1, 21B1, 21C1), Fland alternative key overflow block management table (21A, 21B, 21C), Fland alternative key overflow block management table Pointer (20A1,20B1,20C1) is kept equal to the primary system.

  Access in the accelerator system 3 can be realized by combining the access described using the overflow block management table and the method described in the accelerator system 3.

In the above, explanation was made assuming a symmetric system of accelerator system 3. However, in the case of an asymmetric system, only the accelerator system holding the changed part in the primary system receives the change. Will be updated.

Illustration of the database using the overflow block management table Illustration of location table record Explanatory drawing of record in overflow block management table Illustration of overflow block insertion using the overflow block management table Explanatory diagram of search by alternative key Illustration of variable-length primary block Illustration of reorganization Illustration of reorganization using overflow block management table Diagram explaining the principle of accelerator system Explanatory drawing of application example of overflow block management table of accelerator system

Explanation of symbols

1 ... Primary system 3 ... Accelerator system 10 ... Location table 11 ... Primary block 12 ... Overflow block 13 ... Overflow block 16 ... Fland -Location table 20 ... Overflow-Block management table 22A-Alternative key block ALA0-Alternative key location table ALA1-Fland alternative key location table LC-Current location table LN-New Location table FP ... Last pointer OFP-Overflow-Block management table-Pointer RPLC-Current location table reorganization pointer RPLN-New location table reorganization point Over

Claims (6)

In a data storage and retrieval system for storing and retrieving data in a computer,
A data record with a data field containing a primary key field, and
A primary block that stores data records in primary key order and an overflow block;
A location table record that stores the address of the primary block;
A location table that stores location table records in primary key order;
A final pointer to indicate the end of the used area of the location table;
An overflow block management table for managing the address of the overflow block;
A data storage / retrieval system comprising an overflow block final pointer indicating the end of an area used by an overflow block management table.
In a data storage and retrieval system for storing and retrieving data in a computer,
A data record with a data field containing a primary key field, and
A primary block that stores data records in primary key order and an overflow block;
A location table record that stores the address of the primary block;
A working location table that stores location table records in primary key order;
A working last pointer that indicates the end of the used area of the working location table;
A working overflow block management table for managing the address of the overflow block;
A working overflow block last pointer indicating the end of the used area of the working overflow block management table, and
During reorganization
A working reorganization pointer that indicates the end of the reorganization position of the working location table;
A new location table to store the address of the primary block;
A new reorganization pointer that indicates the end of the reorganization position in the new location table;
A new overflow block management table that manages the address of the overflow block;
A new overflow block last pointer that indicates the end of the used area of the new overflow block management table;
A data storage / retrieval system comprising:
The data storage and retrieval system according to claim 1,
A land location table that stores the address of the primary block;
A data storage / retrieval system, comprising: a overflow overflow block management table for managing an address of an overflow block.
In a data storage and retrieval system for storing and retrieving data in a computer,
An alternate key block and an alternate key overflow block that store alternate key records in the order of the alternate keys;
An alternate key location table record that stores the address of the alternate key block;
An alternate key location table that stores alternate key location table records in order of alternate keys; and
A final pointer to indicate the end of the used area of the alternate key location table;
An alternate key overflow block management table for managing the address of the alternate key overflow block;
A data storage / retrieval system comprising: an alternative key overflow block last pointer indicating the end of a used area of an alternative key overflow block management table.
In a data storage and retrieval system for storing and retrieving data in a computer,
An alternate key block and an alternate key overflow block that store alternate key records in the order of the alternate keys;
An alternate key location table record that stores the address of the alternate key block;
A working alternate key location table that stores alternate key location table records in order of alternate keys; and
A working alternate key end pointer that indicates the end of the used area of the working alternate key location table;
A working alternate key overflow block management table for managing addresses of alternate key overflow blocks;
A working alternate key overflow block last pointer indicating the end of the used area of the working alternate key overflow block management table, and
During reorganization
A working reorganization pointer that indicates the end of the reorganization position in the working alternate key location table;
A new alternate key location table that stores the address of the alternate key block;
A new reorganization pointer that indicates the end of the reorganization position in the new alternate key location table;
A new alternate key overflow block management table that manages the address of the alternate key overflow block;
A new alternate key overflow block last pointer that indicates the end of the used area of the new alternate key overflow block management table, and
A data storage / retrieval system comprising:
The data storage and retrieval system according to claim 4,
The alternate key / location table that stores the address of the alternate key block, and
A data storage / retrieval system comprising: a fland for managing addresses of a substitute key, overflow block, a substitute key, overflow, and a block management table.

Images