JPH0484237A - Update method for overlap data in different type distributed data base - Google Patents

Abstract

PURPOSE:To update overlap data even if an update processing which cannot translate data in a standard data operation language by assembling a data update instruction from position information on an update part and data which is read. CONSTITUTION:At first, the requested update processing is executed in one computer system. Then, it is judged whether data on the updated part exists in the other computer system as overlap data or not. When overlap data does not exist, the processing is terminated. When overlap data exists, it is judged whether the executed update instructed can be translated into the standard data operation language or not. When it can be translated into the standard data operation language, the processing is terminated. When it cannot be translated into the standard data operation language, information on the position of the updated part is obtained and position information is held.

Description

Detailed Description of the Invention (Industrial Application Field) The present invention is a heterogeneous distributed database in which a plurality of 14 different types of computer systems are connected via communication lines, buses, etc., and duplicate data exists between each computer system. Concerning how to update duplicate data in .

(Prior Art) Conventionally, there is a method for updating duplicate data in a multi-model distributed database environment in which multiple computer systems of different models are connected via communication lines, buses, etc., and duplicate data exists between each computer system. Methods have been considered for translating data update instructions executed within one computer system into a standard data manipulation language and transmitting the translated data to other computer systems that have duplicate data.

On the other hand, if a process occurred that could not be translated into a standard data manipulation language, the entire overlap was transferred and replaced with the previous duplicate data.

(Problems to be Solved by the Invention) The conventional method of transferring the entire overlapped portion has the disadvantage that unupdated data in the overlapped portion is also transferred, resulting in unnecessary data transfer costs.

Therefore, an object of the present invention is to update all duplicate data in the same manner when a data update process that cannot be translated into a standard data manipulation language occurs for duplicate data distributed among multiple computer systems of m types. It is an object of the present invention to provide a method for updating duplicate data in a heterogeneous distributed database, which enables updating and further suppresses wasteful data transfer during data update processing.

(Means for Solving the Problems) The method for updating duplicate data in a heterogeneous distributed database of the present invention is a method for updating duplicate data in a heterogeneous distributed database in which duplicate data exists between multiple computer systems of different models connected by communication lines, buses, etc. A method for updating duplicate data in a database, when the duplicate data exists in an updated part by a data update command in a certain computer system, and the data update command cannot be translated into a standard data manipulation language. retaining information indicating a position, reading the contents of the data of the update part based on the position information, and assembling a data update command from the position information and the data of the update part using a standard data manipulation language; The present invention is characterized in that the data update command is transferred to another computer system having the duplicate data.

(Operation) The present invention reads data that has already been updated and assembles a data update command from the position information of the updated portion and the read data, thereby creating an update process that cannot be translated into a standard data manipulation language. It allows updating of duplicate data even if it exists in other computer systems. Furthermore, by retaining information on the location of updated portions, data that has not been updated is not transferred, thereby suppressing wasteful data transfer costs.

(Example) Next, an example of the present invention will be described with reference to the drawings.

FIG. 1 is a diagram showing a flowchart of one embodiment of the present invention. FIG. 2 is a diagram showing an example of the case where the present embodiment is applied to updating a table of a relational database.

The overall processing flow will be explained using FIG.

The processing of this embodiment is broadly divided into two parts.

One is the processing when updating data within a certain computer system, and the other is the processing when instructing another computer system that has data that overlaps with the data updated by that data update to update the duplicate data. be.

First, processing when updating data within one computer system will be explained. First, the requested update process is performed (step 11).

Next, it is determined whether or not the updated data exists as duplicate data in another computer system, and if no duplicate data exists in other computer systems, the process is terminated (Step E2). If duplicate data exists in another computer system, it is determined whether the executed update command can be translated into the standard data manipulation language, and if it can be translated into the standard data manipulation language, the process is terminated (step 13). If it cannot be translated into a standard data manipulation language, information on the location of the updated portion is obtained and retained (step 14).

Next, we will explain the process when instructing another computer system to update duplicate data. 4. Search for the location information of the updated data and recognize the location of the updated part (step 15) .

Next, it is determined whether there is any updated part that has not been processed yet, and if there is no unprocessed part, the processing is terminated (step 16). If there is an unprocessed part, the updated data is The part indicated by the position information is read (step 17).Next, a data update command is assembled from the position information and the value of the read data using the tI data manipulation language (step 18). Then, the assembled update command is transferred to other computer systems that have duplicate data (step 19). When the transfer is completed, the process returns to step 15 to process the next updated portion. In the embodiment shown in FIG. 1, the search for the location information of the updated portion and the assembly and transfer of the update command are performed one by one, but they may be performed all at once. For example, the locations of all updated portions may be recognized first, and instructions may then be assembled. In this way, the overall time required for control when searching for position information is shortened, and optimization such as combining multiple updated parts into one becomes possible.

When transferring update instructions, assembled update instructions may be transferred all at once instead of being transferred one by one. In this way, the number of control messages during transfer can be reduced.

As an example of the processing in steps 14 and 17, an example in which this embodiment is not applied to updating tables in a relational database is shown in Fig. 2. The zero-dot line arrow indicates the flow of processing in step 14 in Fig. 1. , and the solid arrow indicates step 1 in Figure 1.
7 shows the flow of processing in 7. Assuming that the table to be updated has n records with m attributes, prepare an nXm bit matrix.0For example, when the X-th attribute of the y-th record in the table is updated, the bit matrix Set up the and soto at the position corresponding to (x, y).

In step 17, the bit matrix is searched to find position information (x, y), and the attribute name and primary key value are extracted using the correspondence table. For example, if attribute 1 is the primary key item, search the value of attribute 1 in the updated table and read the attribute value of the item whose attribute name is attribute X in record y whose primary key value is py. . The update instructions are then constructed using a standard data manipulation language. For example, SQL
Using language, it looks like this:

UPDATE T BLE (Name of updated table) S
ET attribute x- (read attribute value) 11HERE attribute 1=Py; A data update command is assembled based on the position information of the updated portion in this way.

(Effects of the Invention) As explained above, according to the present invention, redundant data transfer is not performed even when updates of duplicate data existing in other computer systems occur that cannot be translated into a standard data manipulation language. In the case of delayed update (a method of executing the update process after a certain amount of time, such as at night when there is low traffic, instead of updating the data of the @ system immediately) , even if certain data has been updated multiple times, it can be updated to the latest value with a single process.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing a flowchart of an embodiment of the present invention;
FIG. 2 is a diagram showing an example of the case where this embodiment is applied to updating a table of a relational database.

Claims (1)

[Claims] A method for updating duplicate data in a heterogeneous distributed database in which duplicate data exists between multiple computer systems of different models connected by communication lines, buses, etc.
If the duplicate data exists in a part updated by a data update command in a certain computer system and the data update command cannot be translated into a standard data manipulation language, retain information indicating the location of the updated part, and store the position information. reads the content of the data in the update portion based on the data, assembles a data update instruction from the location information and the data in the update portion using a standard data manipulation language, and uses the data update instruction to read the content of the data in the update portion using a standard data manipulation language. A method for updating duplicate data in a heterogeneous distributed database, the method comprising: updating duplicate data in a heterogeneous distributed database.