JPH0477341B2 - - Google Patents

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention provides a method for detecting information that exists in a database when a record definition is changed, such as adding or deleting a data item or changing the attributes of a data item. You can change the definition of the record without reconfiguring the stored record, and access items that match the changed definition (item data specified by the user after adding items, deleting items, and changing attributes). The present invention relates to a method for dynamically changing data items in record-type data, in which data items are accessed and sent back to the user.

Recently, general database systems that have been introduced in various industries such as the financial industry and manufacturing industry tend to become less efficient in accessing the database and in using space as they are used repeatedly. Periodically, download the database,
Normally, a so-called re-editing process is performed, in which the data is initialized and reinserted.

Therefore, there is a need for a database system that does not require such re-editing processing. For example, a database system that allows data items to be added, deleted, and item names changed by trial and error, that is, a database system that can dynamically change data items, is needed. Possible systems are known.

This type of system is characterized by ease of database introduction and operational flexibility.

In addition, the database includes a standardized database (network type database or hierarchical structure type database) that is often used in the banking industry, etc.
There are also so-called non-standard databases (relational databases), which often involve the addition and deletion of data items, and changes in item names, such as those used in personnel management, statistical management, etc. However, in recent industries, With the diversification of data processing operations, there is a need for the above-mentioned amorphous databases, which are easy to introduce and have great operational flexibility, even if their performance is slightly degraded. It is expected that the performance will improve.

[Conventional technology]

As one of the non-standardized database systems mentioned above, an improved method for dynamically changing data items in record-type data was proposed in Japanese Patent Application No. 58-118940.
It is disclosed in.

FIG. 4 and FIG. 5 are diagrams showing the concept of "dynamic change method of data items in record type data" disclosed in the above-mentioned Japanese Patent Application No. 118940/1982,
Figure 4 shows the structure of the record definition body of this method,
FIG. 5 shows the physical structure of records generated by this method.

4 and 5, 1 is a file cell, 2-1 to 2-1' are item cells, 3 is a bit map length section, 4 is a bit map section, and 5-1
5-1' to 5-1' indicate data items, solid line arrows indicate physical order pointers, dotted line arrows indicate display order pointers, and dashed-dotted line arrows indicate new definition pointers. Hereinafter, the same reference numerals indicate the same objects throughout the figures.

An outline of the ``method for dynamically changing data items in record type data'' disclosed in the above-mentioned Japanese Patent Application No. 118940/1988 will be explained below. That is, the bit map length section, the bit map section,
Multiple records with data items (Figure 5,
3, 4, 5-1 to 5-1'), and one or more item cells (see FIG. 4, 2-1') that store the attributes of the data items. 1 to 2-1'), the physical sequence pointer on the record of the data item (see solid line in Figure 4), the record visible from the user interface of the data item, and the data item contained in the record. It has a display order pointer (see dotted line in FIG. 4) provided between item cells to indicate a logical display order indicating the order of
Attribute change data item 5- with new attribute change
Attribute change item cell 2-1' with new attribute definition for 1' and existing item cells 2-1, 2-
2. A record definition storage section configured to provide an attribute change item definition pointer (see FIG. 4, one-dot chain line) between item cells indicating the relationship between ~, the file storage section, and the record definition storage section; and an access system that accesses the data item, and the access system is configured to:
In the case shown in FIG. When the pointer is replaced and deletion of data item 5-2 is commanded {cases in Figures 4 and 5 C}, the corresponding item cell in the record definition storage unit, for example 2-2, is In addition to setting an invalid display, the above display order pointer is changed, and when a command is given to change the attribute of data item 5-1 {cases shown in Figures 4 and 5 D}, the above-mentioned new attribute is retained. Attribute change item cell 2-
1', and communicates between the old attribute item cell 2-1 and the new attribute attribute change item cell 2-1' using the attribute change definition pointer, and the physical arrangement order pointer and display. When the forward pointer is changed and a record input to the file is commanded, the record definition memory has the same number of bits as the number of item cells existing in the record definition storage unit, and the i-th bit is physically of the data item corresponding to the i-th item cell according to the array order pointer,
Indicates the presence or absence in the input target record (0/
1), the data items of the record to be input are arranged according to the contents of the record definition storage section, and the bit map section 4 is written to the arranged data items. In addition, the record is stored in the file storage unit, and when a record is commanded to be output from the file, the corresponding record is retrieved from the file storage unit, and the contents of the bit map unit 4 of the retrieved record are And when there is a necessary process such as adding an item by referring to the record definition storage section, set 'NULL' to the item 5-i.
If there is an item deletion, delete the item 5-j, and if there is an attribute change of item 5-k, change the attribute of the item, Output the processed record.

The above-mentioned processing includes record read processing, which will be described later, mapping processing for recognizing the address of a data item, and item access processing using the recognized address.

Therefore, the above-mentioned improved "dynamic change method for data items in record-type data" is effective when the record definition is changed, such as adding or deleting a data item, changing the attributes of a data item, etc. The feature is that the record definition in the record definition storage section is changed without reconfiguring the stored records existing in the record definition body, and item access that matches the record definition body is performed.

[Problem that the invention seeks to solve]

However, in this method, the record definition changes dynamically as data items are added, deleted, data item attributes changed, etc. Since it is not changed, there is a possibility that each record shown in Figure 5 A to D exists in the file storage unit, and when accessing (reading) a specific record, any of the above A to D Performs the above-mentioned complex record read processing to read the record and match it to the latest definition set in the record definition storage unit, mapping processing that recognizes the address of each data item, and item access processing based on the recognized address. However, there was a problem in that the overhead during file access became large.

The present invention has been developed in view of the above-mentioned drawbacks of the conventional art, and in general, changes to files are not always made, and records that are adjacent to each other are often records that have been changed at the same time. Focusing on the above features, we provide a method that does not require mapping processing to recognize the address of the accessed data item every time a file is accessed (read), and item access processing based on the recognized address. The purpose is to

[Means for solving problems]

The purpose of this is to provide the bitmap length section 3 and the bitmap length section 3 in the record type data format database 17.
A file consisting of a bit map section 4 having a value indicating the presence or absence of the data items 5-1, 5-2, and the number indicated by , and a plurality of records having the data items 5-1, 5-2, and. a file storage unit that stores the data items 5-1, 5-2, and one or more item cells 2-1, 2- that hold the attributes of the data items 5-1, 5-2, and
2. ~, the data item 5-1, 5-2, the physical sequence pointer (solid arrow) on the record of ~, the record visible from the user interface of the data item, and the record contained in the record. It has a display order pointer (dotted line arrow) provided to indicate a logical display order indicating the order of data items, and an attribute change data item 5-1' in which an attribute change has been made to a certain data item. Attribute change item cell 2-1' with new attribute definition
and existing item cells 2-1, 2-2, ~, and the attribute change item definition pointer (dotted chain arrow) between item cells 2-1, 2-2, ~, 2-1'. a record definition storage section configured to provide a record definition storage section; a record access system 12 that accesses the file storage section and the record definition section; When the definition of a record such as adding or deleting ~ or changing the attributes of data items 5-1, 5-2, ~ is changed in the item cells 2-1, 2-2, ~, the above record type data. Without reorganizing the stored records existing in the database 17, only the record definition storage section is changed, and item access (item addition, item addition,
Dynamic modification of data items 5-1, 5-2, ~ in record-type data that accesses the item data specified by the user and returns it to the user after deleting items and changing attributes. In this method, mapping cell 1 is stored in the record definition storage section.
-1 is provided, and the bit map length section 3 of the accessed record is provided in the mapping cell 1-1,
The pattern of the bit map section 4 and the address information for each data item 5-1, 5-2, etc. are stored, and when the record is accessed next, the bit map length of the mapping cell 1-1 is stored. Part 3
Then, by referring to the pattern of the bit map section 4, a mapping process that recognizes the address of the data item in the record only when a mismatch is detected, and a mapping process that recognizes the address of each recognized data item in the corresponding mapping cell. This is achieved by the method for dynamically changing data items in record-type data of the present invention, which includes storing a pattern address in a location, and accessing the item to the corresponding record based on the stored address. .

[Effect]

That is, according to the present invention, the patterns of the bit map length section and bit map section of the last accessed record, and address information such as the record relative address of each data item are stored as mapping cells in the record definition storage section. Then, the next time you access the file, use the above pattern of the record,
The pattern is compared with the content of the above mapping cell, and only when there is a mismatch, the mapping process is performed, and the above pattern of the record is stored as a new mapping cell, and the item access to the record is performed, but if the pattern matches. Sometimes, the address of each data item is obtained from the address information of each data item of the mapping cell, and item access is performed based on the address on the buffer in the physical database system (see 14 in Figure 3). Therefore, by using a method of dynamically changing data items in record-type data, it is possible to obtain access performance that is approximately the same as database access in an unchangeable record format.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the drawings.
FIG. 1 is a diagram showing an example of the configuration of a record definition body when the present invention is implemented, and shows a mapping cell 1-1.
is a record definition body necessary to implement the present invention, FIG. 2 is a flowchart showing the operation when accessing a database by implementing the present invention, and FIG. 3 is a record definition body necessary to implement the present invention. 1 is a diagram showing an example of the configuration of a relational database management system.

Even if the present invention is implemented, the item access operations associated with adding, deleting, changing attributes, inputting records in a file, reading records from a file, etc. are exactly the same as in the conventional method.
The explanation will be omitted here, and will focus on the operation of matching the access pattern of the mapping cell and the corresponding pattern of the record currently being accessed, which is the main focus of the present invention.
Embodiments of the present invention will be described with reference to FIGS. 1 to 3.

First, referring to FIG. 1 and FIG. 2, the operation when reading a record requested by a user from a database will be explained. As is clear from FIG. 1, the present invention differs from the conventional system in that a mapping cell 1-1 is provided in the record definition body.

Step 20: Read the record requested by the user from the database.

Step 21: Read the bit map length part and the bit map part of the record read above, and read the previous access pattern (bit map length part) set in the mapping cell related to the record stored in the record definition storage part. If a matching output is obtained, the process jumps to step 24, but if there is a mismatch, the process moves to the next step.

Step 22: Regarding the read record, follow the display order pointer of item cells 2-1, 2-2, ~, for example, item cell 2 in Fig. 4D.
Bitmap 4 shown in FIG. 5 corresponding to -1
Recognizing that the above display is "O" (NULL) and knowing that the corresponding data item 5-1 does not exist in the record, select the new attribute from the attribute change item definition pointer of item cell 2-1. Find the changed item cell 2-1', confirm that the corresponding bit on the bit map 4 is "1", and check 0/1 of the bit map 4 from the first item cell 2-1 of the physical order pointer. At the same time, add the data length up to the above new attribute change item cell 2-1' to obtain the new attribute change data item 5-1'.
Find the relative data item address within the record. And each data item 5-1, 5 above
-2, the process of outputting the addresses of ~ is processed for all data items, and the process moves on to the next processing step.

Step 23: Generate a bit map length part and a bit map part for the record subjected to the above mapping process, and use the contents as an access pattern and the address information of each data item requested by the above user to create a record definition. Mapping cells related to the record in the storage unit (Fig. 1,
1-1).

Step 24: Based on the record address information stored in the mapping cell or the address information of the user record generated in step 24, only the corresponding data item is returned to the user.

Next, the operation when the database access according to the present invention is executed by the relational database management system shown in FIG. 3 will be explained.

First, an access (read) request from the user 10 to the database 17 is made in a collective interface format. For example, it is an instruction such as "which table, which item, which value is greater than or equal to".

When the relation operation system 11 receives this instruction, it accesses the external schema 15 in which record definitions visible to the user are stored, and converts it into a format that indicates "which item of which record" satisfies the above conditions. and passes it to the record access system 12.

In the record access system 12,
Logical database access
system 13, the logical database access system 13 uses the external schema 15
After accessing and expanding the latest record definition in the main storage (not shown), the expanded record definition information and the physical location of the record in the database 17 are stored. The user request is converted into a form that can be executed by the physical database access system 14 and is passed on by referring to the internal schema 16 that indicates the user request. Thereafter, the information developed on the main storage device is referred to as the record definition body information.

The physical database access system 14 includes:
The record is read by referring to the internal schema 16 and the database 17, and is passed to the logical database access system 13.

Here, the logical database access system 13 converts the record requested by the user into the representation format of the external schema 15 according to the correspondence between the record descriptor information developed in the main storage device and the internal schema 16. (see step 20), extracts the bit map length part and bit map part of the converted record, and inserts the mapping cell 1 in the record definition information.
-1 access pattern and if it matches,
Of the data item addresses stored in the mapping cell, only the data requested by the user is returned to the record access system 12.
(See Steps 21 and 24.) However, if a mismatch is detected in the above comparison process, after performing the above mapping process (see Step 22), the bit map length and bit map length are calculated for the generated record. is the access pattern, and the address information of the record is,
After storing in the above mapping cell (see step 23), only the records requested by the user are
Return to record access system 12.
(See step 24) By returning this to the user 10 via the relation operation system 11, a series of database processing is completed.

When updating the database 17, instructions received by the logical database access system 13 as requests for records in the external schema 15 are updated by the internal schema 16 onto the actual database 17. It is executed in such a way that it is reflected.

Therefore, the above-mentioned mapping process requires a lot of processing time. Therefore, in the present invention, by focusing on the structural characteristics of the database mentioned above, the bitmap length part and bitmap A mapping process that compares the pattern of the data item with that of the record at the time of previous access (read) and recognizes the address of each data item only if it is different, and stores the recognized address in the corresponding area of the bit map cell. However, the feature is that item access is performed based on the stored address.

〔Effect of the invention〕

As explained above in detail, the method for dynamically changing data items in record-type data of the present invention stores the patterns of the bit map length part and bit map part of the last accessed record in the record definition storage part. It is stored as a mapping cell, and the next time the file is accessed, the above pattern of the record is compared with the contents of the mapping cell, and only when there is a mismatch, the mapping process is performed and the above pattern of the record is updated. The data item is stored as a mapped cell and the item is accessed for the record, but when the pattern matches, the item is accessed according to the contents of the mapped cell, so it is not necessary to change data items in record type data. By using a method of dynamically performing this, it is possible to obtain almost the same access performance as database access in an immutable record format.

[Brief explanation of the drawing]

Fig. 1 is a diagram showing an example of the configuration of a record definition body when the present invention is implemented, Fig. 2 is a flowchart showing the operation when accessing a database by implementing the present invention, and Fig. 3 4 is a diagram showing a configuration example of a relational database management system to which the present invention is applied, FIG. 4 is a diagram illustrating the configuration of a record definition body in the conventional method, and FIG. 5 is a diagram showing a record generated by the conventional method. FIG. 2 is a diagram showing the physical configuration of. In the drawings, 1 is a file cell, 2-1 to 2
-1' is an item cell, 1-1 is a mapping cell, 3
is the bit map length section, 4 is the bit map section,
5-1 to 5-1' are data items, 10 is a user,
11 is a relation operation system, 12 is a record access system, 13 is a logical database access system, 14 is a physical database access system, 15 is an external schema, 16
is internal schema, 17 is database, 20-2
4 indicates operation steps.

Claims (1)

[Scope of Claims] 1. In the database 17 in record type data format, a bitmap length section 3;
A file consisting of a bit map section 4 having a value indicating the presence or absence of the data items 5-1, 5-2, and the number indicated by , and a plurality of records having the data items 5-1, 5-2, and. a file storage unit that stores the data items 5-1, 5-2, and one or more item cells 2-1, 2- that hold the attributes of the data items 5-1, 5-2, and
2. ~, the data item 5-1, 5-2, the physical sequence pointer (solid arrow) on the record of ~, the record visible from the user interface of the data item, and the record contained in the record. It has a display order pointer (dotted line arrow) provided to indicate a logical display order indicating the order of data items, and an attribute change data item 5-1' in which an attribute change has been made to a certain data item. Attribute change item cell 2-1' with new attribute definition
and existing item cells 2-1, 2-2, ~, and the attribute change item definition pointer (dotted chain arrow) between item cells 2-1, 2-2, ~, 2-1'. a record definition storage section configured to provide a record definition storage section; a record access system 12 that accesses the file storage section and the record definition section; When the definition of a record such as adding or deleting ~ or changing the attributes of data items 5-1, 5-2, ~ is changed in the item cells 2-1, 2-2, ~, the above record type data. By changing only the record definition section described above without reconfiguring the stored records existing in the database 17,
In a method for dynamically changing data items 5-1, 5-2, etc. in record-type data in which item access is performed in accordance with the changed definition structure, mapping cell 1 is stored in the record definition storage section.
-1 is provided, and the bit map length section 3 of the accessed record is provided in the mapping cell 1-1,
The pattern of the bit map section 4 and the address information for each data item 5-1, 5-2, etc. are stored, and when the record is accessed next, the bit map length of the mapping cell 1-1 is stored. Part 3
Then, a mapping process (processing step 22) that refers to the pattern of the bit map section 4 and recognizes the address of each data item constituting the record only when a mismatch is detected; storage in the corresponding area of the mapping cell 1-1 (processing step 23), and item access based on the recognized address (processing step 24)
A method for dynamically changing data items in record type data.