JPH05289914A - Data base control system - Google Patents

Abstract

PURPOSE:To automate the start of reorganization by executing reorganization without stopping the use of a data base and setting up the upper limit of a storing range. CONSTITUTION:This data base control system is constituted of an allowable value setting means 1 for setting up the storing ranges of respective records set up by a user in an allowable value management table 3 while referring to definition information 2, a storage status monitoring means 4 for obtaining data base updating information from a DBMS 7 for retrieving/updating the data base 8 and controlling the rearrangement of records while comparing the records with the contents of the table 3 and a record rearranging means 5 started from the means 4 to rearrange records, and when a storage state is not improved even after rearranging the records, output an alarm report 6.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention connects a logical relationship (set) between records stored in an auxiliary storage device and having a plurality of formats in a chain using a pointer using the storage address of the record. The present invention relates to a database control system for controlling a database represented by doing.

[0002]

2. Description of the Related Art A logical relationship (set) between records stored in an auxiliary storage device and having a plurality of formats is expressed by connecting them in a chain using a pointer using a storage address of the record. In a given database management system, a set (cluster) of a record (parent record) and a record (child record) connected to the record in a set relationship generally spans multiple blocks (units of physical input / output). Stored.

The user of the database constructs the system by predicting the number of blocks stored in each cluster from the expected value of the number of records actually stored. However, actually, even in the database designed in this way, the storage state may be disturbed as a result of addition or deletion of records due to long-term operation.

As a result, the total number of blocks stored in a certain cluster becomes larger than originally planned, and when the database is accessed, the physical input / output occurs more times than planned and the system performance is degraded. It will appear. When such system performance degradation occurs, reorganization processing for correcting the storage state of the database is performed.

In the conventional method, the reorganization timing is checked by actually executing a program for measuring the storage state, whether or not the processing performance actually deteriorates. Moreover, the processing method of reorganization was a batch processing. This is a method in which all the records in the database are saved to another file and the database is initialized, and then the saved records are written back to the database in the order in which they are linked as a set.

[0006]

In the above-mentioned conventional reorganization method, since the records are written back in a batch after the database is saved to another file, the user selects the database to be reorganized. There was a problem that all the business programs to access had to be stopped. Further, the present invention for performing reorganization is made in order to solve the problems in the conventional example as described above, and enables reorganization without stopping the operation of the database, and the upper limit of the storage range is set. It is an object of the present invention to provide a database control system that can automatically start reorganization by setting it.

[0007]

In order to achieve the above object, the database control system according to the present invention sets a set of records having a plurality of formats stored in an auxiliary storage device as a storage address of the record. In a database that is expressed by connecting in a chain using a pointer using, a logical pointer that represents the logical appearance order of records in the above set within one record, and a physical appearance order in the set In a database management system that has a physical pointer that represents and that performs a reconnecting process of the physical pointers as needed in addition to the update process similar to the conventional process when updating a record
Calculated from the average number of records contained in one set 1
A permissible value setting means for managing the storage range allowed for records belonging to one set, and a physical pointer for the storage range of the records connected by each logical pointer based on the storage range allowable value set by the above-mentioned permissible value setting means. A storage status monitoring means for monitoring the storage range by using, and calling the record rearrangement means when necessary,
And a record rearrangement unit which is activated by the storage state monitoring unit and rearranges the target record group when the storage range of records belonging to a certain set exceeds the value set by the allowable value setting unit. And

[0008]

In the present invention, the storage state monitoring means uses the physical pointer to determine the storage range of the record connected by each logical pointer based on the storage range tolerance set by the tolerance setting means. The record relocation means is called as necessary. The record rearrangement unit is activated by the storage state monitoring unit to rearrange the target record group when the storage range of the records belonging to a certain set exceeds the value set by the allowable value setting unit.

[0009]

Embodiments of the present invention will now be described with reference to the drawings.

FIG. 1 is a conceptual diagram showing the relationship between records and sets. Circles represent records and line segments represent sets. In FIG. 1, the record a corresponds to a parent record and the record b corresponds to a child record.

FIG. 2 is a conceptual diagram showing the relationship between the logical pointer and the physical pointer. Child records are connected under the parent record a1 in the order of b1, b2, b3, b4, b6, b5, and this is a logical pointer indicating the logical relationship between the records. On the other hand, from a1 to b1, b
2, b3, b4, b5, b6 and the records are connected in the order of the storage address is the physical pointer. In both pointers, the parent record has the addresses of the first and last records of the child records.

In the example of FIG. 2, the logical pointers are b5 and b.
1. The physical pointers correspond to the addresses b6 and b1. The first child record has the address of the parent record instead of the address of the immediately preceding child record, and the last child record has the address of the parent record as a pointer instead of the address of the immediately following child record.

Further, FIG. 3 is a conceptual diagram of the control information in the record. Each record consists of a record header part, a logical pointer part, a physical pointer part and the contents of the record.

FIG. 4 is a configuration diagram for implementing the present invention. According to the present invention, referring to the definition information 2, the allowable value setting means 1 for setting the allowable value of the storage range of each record set by the user in the allowable value management table 3, and the DBMS 7 for searching / updating the database 8 The storage status monitoring means 4 for controlling the rearrangement of records while obtaining information for updating the database and comparing the contents of the allowable value management table 3, and the storage status monitoring means 4 are activated to rearrange the records, Of the record rearrangement means 5 for outputting the warning report 6 when the storage state is not improved even if the rearrangement is performed.

FIG. 5 is a diagram showing an example of a storage range allowable value management table 3 used in the present invention. This figure shows that even if the record name of the parent record of the cluster whose storage range is to be managed, the set name, the maximum number of blocks in the storage range corresponding to each, and the actual storage range exceed the storage allowable value. It has the information of the control SW for controlling not to rearrange the cluster.

FIG. 6 is a diagram showing the storage states before and after the rearrangement of records. a) corresponds to the storage state before the rearrangement, b) corresponds to the storage state after the rearrangement, and c) corresponds to the storage state after the rearrangement. Square 1 in this figure
Each one corresponds to one block in the database, circles correspond to records, and curves connecting the records correspond to sets. Also,
The shaded blocks are work areas used for rearrangement.

FIG. 7 is a diagram for explaining a procedure for actually rearranging records by taking a certain cluster. As in FIG. 6, one square corresponds to one block in the database, circles correspond to records, and curves connecting the circles correspond to sets. The example taken here is a database that can store eight records in each block.

FIG. 6 shows the initial state when a) is rearranged.
B) is extracted from one cluster.
The a) is an initial state when the records are rearranged, and is located in the record a1 in the first layer of the cluster.
b) to d) show that the record of the second layer of the corresponding cluster is being moved. In e), the rearrangement up to the second layer is completed, which is the first step of moving the third layer. In f), when one record in the third layer is moved, the same processing can be applied to the remaining records. If so, the rearrangement process for this cluster ends.

FIG. 8 is a diagram showing a procedure in which the allowable value setting means 1 for managing the storage range of records updates the allowable value management table 3. When the user requests the update of the allowable value management table 3, the allowable value setting means 1 is activated. The tolerance value setting means 1 receives information on the tolerance value that the user is trying to set (step 801).

It is judged by referring to the definition information 2 of the database 8 whether or not the received information is correct, for example, the relation between the record and the set is incorrect (step 802), and if the allowable value to be set is incorrect. If so, inform the user that the input information is incorrect (step 80
9) End the process.

When there is no error in the information set by the user, the allowable value setting means 1 searches the allowable value management table 3 to check whether the same record / set combination has already been registered. Yes (Step 80)
3). If the same combination does not exist, the input information is added to the allowable value management table 3 (step 808) and the process ends.

When the same record / set combination is already registered in the table, the control switch ON / OFF is checked for the target combination (step 804). If it is OFF, the information is set in the allowable value management table 3 and the process ends. When the control SW is ON, the storage range of the allowable value is checked next (step 80).
5).

If the newly set storage range is wider than the conventional storage range, the control SW is turned off (step 80).
6) If it is narrow, a report warning that a cluster exceeding the storage range to be set already exists is output, and then the allowable value management table 3 is updated (step 808) and the process ends.

FIG. 9 is a diagram showing an outline of the operation of the storage state monitoring means 4 which monitors the storage state of each record and activates the record rearrangement means 5 when necessary.

The storage state monitoring means 4 monitors the update of the database 8, and when the update of the database 8 occurs (step 901), the following processing is performed.

First, the allowable value management table 3 is searched to check whether or not the allowable value of the storage range is set for the combination of the record to be updated and the set (step 902). If the allowable value is not set, the process ends without doing anything. If there is an allowable value, the cluster to which the updated record belongs is specified from the update information, and the storage range of the relevant cluster is determined from the physical pointer of the parent record (step 903).

Next, the allowable value of the storage range corresponding to the target cluster is read from the allowable value management table 3 (step 90).
4) Then, the actual storage range is compared with the allowable value of the storage range (step 905). If the storage range is within the allowable value, the following processing is skipped and the processing ends.

Subsequently, the record rearrangement control SW O
N and OFF are checked (step 906), and if ON, record rearrangement is not performed and processing ends. If the control SW is OFF, the record rearrangement means is called (step 90
7) The process after the record rearrangement is completed.

FIG. 10 is a diagram showing an outline of the operation of the record rearrangement means 5.

The record rearrangement means 5 activated by the storage state monitoring means 4 first determines the size of the work area required for the rearrangement based on the record storage range (step 1001) and secures the area. (Step 1002).

Next, a process of moving all the records except the records of the first layer to the secured work area is performed for all the records in the block storing the cluster for the record rearrangement. (Step 100
3). The state in which this processing is performed corresponds to b) in FIG.

Next, the record serving as the entry point of the cluster to be relocated and the record at the top of the cluster are positioned (step 1004), and each record is relocated to the cluster having this record as a parent record. Is performed (step 1005). Details of the record rearrangement process for one designated cluster will be given in the following description of FIG.

When the storage area exceeds the allowable value and the rearrangement of the cluster that is the rearrangement target is completed, the rearrangement is performed for all other records in the rearrangement target block. For this purpose, first, among the records other than the relocation target, the record at the top of some set is searched in order (step 1006), and it is checked whether or not the record exists (step 1007). If exists, the record is rearranged with respect to the cluster having the corresponding record as a parent (step 1008).
After this, in order to check the presence / absence of another record, the process is repeated from step 1006.

If there is no record in the check in step 1007, it is rechecked whether or not the cluster to be rearranged is stored within the range indicated by the allowable value (steps 1009 and 1010). If the data can be stored within the allowable value range, the rearrangement control SW of the allowable value management table 3 is turned off (step 1011) and the process is terminated. On the other hand, if the cluster is not completely within the allowable value of the storage range even after performing the rearrangement, the rearrangement control SW of the allowable value management table 3 is turned on (step 101
2) After outputting a report warning that the storage range exceeds the allowable value even after relocation (step 1013),
The process ends.

The storage range is out of the allowable value range even after the rearrangement is performed, that is, the relevant cluster is in the best storage state (only the records of that cluster are stored under the conditions of the database at that time). , Even if it is packed from the beginning of the block without a gap), it means that the records have already been stored for the cluster in excess of the design limit when the allowable value of the storage range is exceeded. is doing.

In other words, regarding this part, the database does not satisfy the performance at the time of designing, and in such a state, after redesigning the database, reorganization by collective processing,
It will not be resolved unless reconfigured. If this situation occurs, it cannot be solved by the method of the present invention, and if it is left as it is, the relocation processing of records will be uselessly performed every time the problem cluster is updated. For this reason,
Even if the relocation is performed once, if the storage range does not fall within the permissible value, the switch for controlling the relocation suppresses the relocation processing of the problem cluster.

FIG. 11 is a diagram showing the details of the processing for actually rearranging the records.

First, the positioned record is returned from the work area to the database body side (step 110).
1). Records that have been positioned since
We will call it the current record.

Next, it is checked whether or not there is a set whose parent is the current record (step 1102). if,
If such a set exists, the first child record belonging to that set is searched (step 1103).
Check for child records (step 1104),
If there is a child record, the child record is set as the current record, and the record rearrangement process (process of FIG. 11) is recursively called (step 1105).

Here, the record rearrangement process is executed,
Complete the relocation of the cluster whose parent is the currently current record. In order to perform this process on the clusters related to all the sets having the current record as a parent, the processes of steps 1102 to 1105 are repeated.

If the corresponding set does not exist in step 1102, a record at the same level as the current record (child record connected by the same set) is searched (step 1106) and if the record exists. If (step 1107), the process is repeated from step 1101. If there is no record at the same level as the current record in step 1107, the record rearrangement process is terminated.

The operation of the record rearrangement of FIG. 11 will be described in detail using the example of FIG. First, the initial state is a)
Thus, the highest record a1 of the rearrangement target cluster is the current record. From this state, the record to be rearranged is searched. First, the set s1 connected to a1 is traced to make the next record b1 the current record (b in FIG. 7)).

In this state, the record rearrangement process of FIG. 11 is performed. First, the current record b1 is transferred to the database body (c in FIG. 7)). Then, it is checked whether or not there is a child record connected to b1 (steps 1102 to 1102).
1104). In the case of b1, there is no child record, so the process of moving records b1 and below (step 1105) is skipped, and the next record b2 linked by the set s1 is searched (steps 1106 to 1107).
b2 has no child record like b1, so b
Only move 2 to finish.

Next, processing is performed with b3 as the current (d in FIG. 7). First, b3 is returned to the database body and a child record connected to the set s2 is searched.
Since the child record exists in b3, the record rearrangement operation is performed using the first child record c1 of the set s2 as the current record (step 1105). Since the child records of b3 and below have no further lower records, the operations of steps 1103 to 1105 are skipped and only the processing of moving the records of c1 to c5 is performed.

FIG. 7f shows a state in which only c1 is returned to the database main body, the current record is moved to c2, and c2 itself is moved and the presence or absence of child records below c2 is checked. ing. c5
When the move is completed, the record of the same level as c1 to c5 does not exist anymore, so the relocation processing of the record called in step 1105 ends. Up to this point, b
The rearrangement of the records of 3 or less is completed, and the record of the same level as b3 is searched next. However, since this record does not exist, the processing ends, and the rearrangement of clusters connected from a1 ends.

Since eight records can be stored in one block in the example of FIG. 7, the cluster taken as an example can be stored in four blocks before rearrangement, but in two blocks, it will be sufficient. The storage state has been improved by relocation.

[0047]

As described above, according to the present invention, the user's business program can freely update and search, except for the block for which the reallocation processing is performed, so that the user can perform database processing without stopping the business processing. The effect of being able to correct the disorder of the storage state by performing the reorganization of. In addition, the database control system determines when to actually start reorganization, and automatically reorganizes the database. Therefore, the database storage status is always maintained in a good state, and the performance of business programs is not degraded. There is also the effect that there is no. Furthermore, in addition to the automation of reorganization, even if the present invention cannot perform reorganization, a warning report to that effect is output, so the user should measure the storage range in order to know the time of reconstruction. There is also an effect that it is not necessary to perform special processing.

[Brief description of drawings]

FIG. 1 is a conceptual diagram showing a relationship between a parent record, a child record, and a set.

FIG. 2 is a conceptual diagram showing correspondence between a logical pointer and a physical pointer.

FIG. 3 is a conceptual diagram showing control information in a record.

FIG. 4 is a configuration diagram showing an embodiment of the present invention.

FIG. 5 is an explanatory diagram showing an example of an allowable value management table.

FIG. 6 is a schematic diagram showing a state of each record when the records are rearranged by the method of the present invention.

FIG. 7 is a schematic diagram showing in more detail the procedure for extracting and rearranging one cluster in FIG. 6;

FIG. 8 is a flowchart showing an outline of operation of an allowable value setting means.

FIG. 9 is a flowchart showing an outline of operation of a storage status monitoring unit.

FIG. 10 is a flowchart showing an outline of operation of a record rearrangement unit.

FIG. 11 is a flowchart showing an operation of relocation of records.

[Explanation of symbols]

 1 Allowable value setting means 2 Definition information 3 Allowable value management table 4 Storage status monitoring means 5 Record rearrangement means 6 Warning report 7 DBMS 8 Database

Claims (1)

[Claims] 1. A database that expresses a set between records stored in an auxiliary storage device and having a plurality of formats by connecting them in a chain using a pointer using the storage address of the record, One record has a logical pointer indicating the logical appearance order of the records in the set and a physical pointer indicating the physical appearance order of the set, and the physical pointer is added as necessary in addition to the update processing when updating the record. In a database management system that performs pointer reconnection processing, an allowable value setting means that manages a storage range allowed for records belonging to one set calculated from the average number of records included in one set predicted at the time of designing a database. And the record connected by each logical pointer based on the allowable value of the storage range set by the allowable value setting means. The storage range of the record is monitored by determining the storage range by using the physical pointer, and the storage state monitoring means for calling the record rearrangement means as necessary, and the storage range of the record belonging to a certain set are allowed above. A database control system comprising: a record rearrangement unit that is activated by the storage state monitoring unit to rearrange a target record group when the value set by the value setting unit is exceeded.