JP2629072B2 - Distributed database system and data retrieval processing method thereof - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distributed database system and a data search processing method thereof, and more particularly, to a public network in which a plurality of terminals each having a distributed individual database are connected. The present invention relates to a distributed database system and a data search processing method thereof.

(Background technology)

FIG. 16 shows a configuration example of a distributed database system. 5 is a public network, 6, 7 and 8 are terminals connected to the public network 5, 6a, 7a and 8a are databases of terminals 6, 7 and 8, respectively, 6b, 7b and 8b are terminals 6, 7 and 8 8 is a table of contents of the distributed database system included in each of FIGS.

Each terminal 6, 7 and 8 has a database 6a, 7a and
Have 8a and maintain its table of contents 6b, 7b and 8b. The terminals 6, 7, and 8 use the table of contents to search for a terminal that manages the data in response to the data search request, access the terminal via the public network 5, and request and retrieve the search data.

The case where the searched data is the display screen will be described in detail.

In a distributed database, a terminal such as a workstation performs data processing while interacting with a display screen.

The terminal performs a process of searching for and displaying a display screen required by the user. The search processing of the display screen needs to be configured to be executed at a high speed in order to increase the efficiency of data processing.

In order to search for a display screen registered in the terminal, a configuration is adopted in which a keyword assigned to a display screen name is specified. That is, the terminal is provided with a keyword table for managing one or a plurality of keywords assigned to each display screen name in units of display screen names as shown in FIG. When the user specifies a keyword, the terminal displays a display screen to which the keyword specified according to the keyword table is assigned (a list is displayed when there are a plurality of keywords). From the above,
The search process of the display screen is being executed.

The keyword information assigned to the display screen of each terminal is not managed at all according to the concept of distributed processing.

However, if a keyword is specified, one terminal
Executes a search procedure that lists the display screen names to which the specified keyword is assigned by searching whether the specified keyword is assigned in order from the display screen name at the top of the keyword table. Must. That is, the entire keyword table must be searched. Therefore, the search time of the display screen becomes long. Further, the search time becomes longer according to the number of registered display screens.

In addition, since the keyword information assigned to the display screen of each terminal is not managed, the display screen must be searched for all the terminals in accordance with the keyword. Therefore, the display screen of the terminal of the same distributed database system cannot be easily searched.

Further, the load of the data search as described above increases for a single terminal, so that the original processing may not be performed.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a distributed database system capable of searching data registered in a terminal as a distributed database at high speed, and a data search processing method thereof.

Another object of the present invention is to provide a distributed database system capable of searching a display screen registered in a terminal at high speed and a data search processing method thereof.

Further, the present invention provides a distributed database system in which predetermined data among data registered in a plurality of terminals as a distributed database is managed by one predetermined terminal, and a data search processing method therefor. With the goal.

[Disclosure of the Invention]

In the present invention, a predetermined terminal is a temporary center, and data search information such as keywords and table of contents information on data that may be shared in the database of each terminal is registered in the temporary center. The temporary center performs a process of searching for a terminal that manages the search data. As a result, the load on each terminal in the data search process can be reduced.

In the present invention, the data search information registered in the temporary center is stored in a keyword link file.
The keyword link file stores the keyword of the data and at least the number of the terminal that manages the data, that is, the keyword, as a set. Thus, the terminal that manages the data can be easily specified. Therefore, it is possible to quickly find out which terminal should perform the search processing, so that the search for the data can be performed extremely fast.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention, FIG. 2 is a diagram illustrating an embodiment, FIG. 3 is a diagram illustrating a component file and a keyword file, FIG. 4 is a schematic diagram of a keyword table of contents, and FIG. FIG. 6, FIG. 6 is a diagram for explaining the relationship between records, FIG. 7 is a diagram showing an example of matching link data, FIG. 8 is a schematic diagram of a keyword link file, FIG. 10 is an explanatory diagram of the center system, FIG. 11 is an explanatory diagram of communication between the center system and the terminal, FIG. 12 is a flowchart of a data version check process, FIG. 13 is a registration process flowchart, and FIG. 14 is an update process Flowchart, FIG. 15 is a search processing flowchart,
FIG. 16 is a block diagram of a conventional distributed database system.
FIG. 17 is a configuration diagram of a conventional keyword table.

[Best mode for carrying out the invention]

FIG. 1 is a block diagram showing the principle of the present invention, showing a distributed database system according to the present invention.

Reference numeral 1 denotes a terminal to which the present invention is applied, which manages data in a distributed database, for example, registered display screens. Reference numeral 3 denotes a terminal to which the present invention is applied, which is a temporary center (hereinafter, referred to as a center system) and manages a plurality of terminals 1 for predetermined data, for example, a registered display screen. . Reference numeral 10 denotes a man-machine interface provided in the terminal 1, reference numeral 11 denotes an input / output control unit of the terminal 1, which executes input / output control with the man-machine interface 10, and reference numeral 12 denotes a command allocation unit. To execute command assignment processing that specifies the attributes of the display screen, 13
Is a component file for managing attribute information such as commands and keywords assigned to registered display screens, and 14 is a keyword file pointed to by the component file 13 and set on the display screen. One for managing one or more keywords in units of command names, 15 is a link file registration unit, which is a link file provided in the center system 3.
A keyword file is registered for 32, a table of contents file 16 manages command names to which each keyword is assigned in units of keywords, and a table of contents file registration unit 17 is a keyword file 14 Is a communication control unit for executing data communication with another terminal 1 or the center system 3 via the public network 4, and 30 is a center system. 3 is a man-machine interface, and 31 is an input / output control unit of the center system 3.
A keyword link file 32 for executing an input / output control unit with the interface 30 manages the terminal numbers of the terminals 1 in which each keyword is registered in units of keywords. A communication control unit 33 that also manages the associated command names, executes data communication with the terminal 1 via the public network 4.

When there is a display screen registration request from the man-machine interface 10, the input / output control unit 11
A process of registering attribute information such as a name, a command name, and a keyword for a display screen for which a registration request has been made is executed with respect to 13. At this time, the command allocating unit 12 performs a command name allocating process. The actual data of the keyword associated with the command name is stored in the keyword file.
Registered at 14. When data of a new command name / keyword pair is registered in the keyword file 14, the link file registration unit 15 stores the registered keyword (preferably in addition to this) in the link file 32 of the center system 3. Command data) and its own terminal number (preferably, in addition to this, a command name). On the other hand, the table of contents file registration unit 17 stores the paired data of the registered command name and keyword in the table of contents file 16.
Register with. In practice, the terminal number for managing the command can be specified from the command name. That is, the command or the command name is assigned to the data in the database, and is configured to have a data structure that is one of the attributes of the data and that can specify the terminal that manages the data. It is preferable that the terminal which actually manages the data from the command name has a data structure which can directly or indirectly specify the terminal (of course, the terminal which cannot specify the terminal by the command name itself). In some cases, the terminal is specified by a combination of the terminal number and the command name.)

From the man-machine man interface 10 on the terminal 1 side,
When a search request for a display screen is made by specifying a keyword, the input / output control unit 11 executes processing for searching not the keyword file 14 but the table of contents file 16. The searched table of contents file 16 manages command names in units of keywords. Therefore, the input / output control unit 11
Can specify a command name required by one access process. Therefore, by referring to the component file 13, a display screen name corresponding to the specified command name is specified, and the name of the display screen is specified. To display on the man-machine interface 10 a display screen associated with. When a plurality of command names are managed in the table of contents file 16, the names of the display screens corresponding to the respective commands are displayed on the display screen of the man-machine interface 10. Thus, according to the present invention, the corresponding command name can be specified by accessing the table of contents file 16 only once, so that the display screen can be displayed.

The terminal 1 searches the link file 32 of the center system 3 via the communication control unit 18 when there is no command name corresponding to the keyword requested to be searched. By the search processing of the link file 32, the terminal number that manages the keyword requested to be searched is specified. And the terminal 1
The terminal 1 of the specified terminal number via the communication control unit 18
, A keyword is transmitted to execute a display screen search request, and the display screen searched according to the above-described processing procedure is transferred. As described above, the corresponding display screen is displayed. At this time, when the link file 32 manages the command name, by transmitting this command name to the terminal 1 of the specified terminal number, the search for the display screen to be transferred immediately is executed. As described above, according to the present invention, it is possible to know which terminal 1 should execute the search processing according to the center system 3, and therefore, the search of the display screen can be executed at extremely high speed.

FIG. 2 illustrates the configuration of the terminal 1 and the center system 3 to which the present invention is applied. In the figure, the same components as those described in FIG. 1 are indicated by the same symbols. 10a is a display corresponding to the man-machine interface 10, 11a is an input / output table processing unit corresponding to the input / output control unit 11, and 19 is a command specifying unit, which is read from the table of contents file 16 (particularly, a keyword table of contents file 16c). A command for specifying one or more of the command names in accordance with the interaction processing with the operator, 20 is an actual data file,
Data such as files with display screens (hereinafter referred to as
A component actualization table 21 manages the correspondence between the command name assigned to the component and the storage address of the component in the actual data file 20.

The table of contents file 16 contains three files 16a, 16b and 1
Consists of 6c. That is, a keyword table of contents index management file 16a that manages the keywords of the keyword file 14 as an index,
A keyword table of contents index file 16b, which manages the keywords in the keyword file 14 by assigning item numbers
And a keyword index file 16c to which link data and matching link data are added by the index file 16b.

Although the final keyword index file 16c can be created from the keyword file 14, it is possible to register and search the file more quickly by using the keyword index index file 16b as described above. it can.

In accordance with the structure of the table of contents file 16 as described above, the link file 32 of the center system 3 also has three files.
It consists of 32a, 32b and 32c. That is, a keyword link index management file 32a and a keyword link index file 32b are provided to manage the keyword link file 32c. These center systems 3
The structure of each file 32a, 32b and 32c on the terminal 1 side is basically the same as the structure of each file 16a, 16b and 16c on the terminal 1 side.
In this embodiment, a communication control mechanism between the terminals 1 and a communication control mechanism between the terminals 1 and the center system 3 that manages the terminals 1 are omitted for convenience of drawing.

FIG. 3 shows a detailed embodiment configuration of the component file 13 and the keyword file 14. As shown in this figure, the component file 13 is a component name registered in the actual data file 20, a screen name of a display screen, a comment, a keyword link (abbreviated as a KW link in the figure), a command name, and a version number. , Flags and the like. The item of the screen name stores the name of the screen described in a length of up to 40 characters and the abbreviation of the screen described in a length of up to 10 characters. The keyword link points to the position in the keyword file 14 of the command name assigned to the screen name. Here, the command name assigned to a component describes the behavior (that is, processing) of the component. The version number indicates how many times the component having the screen name, that is, data, has been revised. Therefore, the one with the larger version number is the latest component immediately. The flag indicates the presence / absence of a component having the screen name, that is, data external registration. For example, when the flag is “0”, the component is not allowed to be referenced at all except the terminal 1. Therefore, each file in the own terminal 1 has information about the component, but information about the component does not exist in the center system 3 and each file in the other terminal 1. That is, the part is not open. On the other hand, if the flag is “1”, the part is open. Therefore, information on the component exists in a predetermined format of each file in the own terminal 1, the center system 3 and, if necessary, the other terminal 1.

The keyword file 14 is configured to manage one or more keywords KW1, KW2,... Assigned to the command name of each component, pointed to by the keyword link of the component file 13.

FIG. 4 shows a table of contents file 16 which is one of the features of the present invention.
Is illustrated. By using the keyword index file 16c, the index file 16 manages the keywords assigned to each command name in the unit of a command name by using the keyword index file 16c. To manage the command name to which each keyword is assigned.
The registration (creation) processing of the table of contents file 16 by the rearrangement or the like is executed by the table of contents file registration unit 17.

Fig. 5 shows the three files 16 that make up the table of contents file 16
Details of a, 16b and 16c are shown.

The keyword index index management file 16a is composed of the first characters [A],
, Etc., and the starting item number of the first character thereof are managed.

Here, item number is represented by 7-digit number of X ₁ to X _7.
The upper two digits “X ₁ X ₂ ” of the item number are used as the identifier of the first character of the keyword. For example, "01" is "A", "0"
"2" represents "i" and "51" represents "A". The remaining 5 digits of the item number “X ₃ X ₄ X ₅ X ₆ X ₇ ” are values from 00001 to 99999. The item number is assigned to each record in the table of contents file 16 and is uniquely determined.

The starting item number for the first character "A" is "0100000" in the keyword table of contents index management file 16a.
It is said. That is, it is an identifier.

The keyword table of contents index file 16b registers the inputted keywords with a more detailed item number based on the item number of the first character registered in the keyword table of contents index management file 16a, for example, in the order of the Japanese syllabary. However, a newly input partway through is registered in the rear row of the file 16b. The numbers are rearranged in the order of the Japanese syllabary every certain period. For example, the keywords "Oh" and "Ai"
Are “0100010” and “0100020”, respectively.
The identifiers are equal. Further, the keyword “Ai” additionally registered later is set as the item number “0100011” between the former two item numbers, and is registered at the back of the file 16b.

The keyword table of contents 16c basically manages keywords and command names (hereinafter also referred to as commands), and also manages item numbers. The item number is a number assigned to each keyword, and is registered as a management number at the beginning of a record in the index file 16b. Item numbers are rearranged every certain period in the same way as in the index file 16b to arrange the Japanese syllabary numbers. Therefore, referring to the index file 16b, it is possible to determine which keyword follows a certain keyword (record). It is for management. The item number is for linking records to increase the number of commands when the number of commands exceeds one record. For example, as for the keyword “Ah”, first, a corresponding command is stored. Also, the own item number “0100010” is stored as the item number. As the item number, the following keyword "Ai"
The item number “0100011” is stored. This makes it easy to search for a keyword in the order of the Japanese syllabary, regardless of the physical position on the file, when the keyword “Ai” is newly input. The item number is “Nu11”. This indicates that there is no next record for the keyword "Oh". On the other hand, regarding the item number of the keyword “Ai”, “0100021” is stored. In the item number “0100021”, a command corresponding to the keyword “Ai” and not stored in the record of the item number “0100020” is stored. Therefore, the keyword “Ai” has two item numbers.

As described above, the terminal 1 searches for the desired display screen from the registered display screens according to the keyword by providing the table of contents file 16, particularly the keyword table of contents file 16c, and searching for it using the keyword. As you go, you can quickly find the display screen you need.

Also, a keyword table of contents index management file 16a that manages the first character and the item number associated therewith,
By providing a keyword table of contents index file 16b that manages keywords according to the item numbers of this file 16a, a quick search using keywords becomes possible.

Furthermore, a keyword table of contents file 16c rearranged in the order of predetermined characters can be created by the management file 16a and the table of contents index file 16b, whereby a quick search can be performed in the order of the predetermined characters. Furthermore, since the keyword index file 16c is registered in the predetermined character order using the keyword index index management file 16a and the keyword index index file 16b, the search in the predetermined character order can be easily performed.

In addition, since the keyword index file 16c is created after the item number paired with the keyword is once prepared in the keyword index file 16b, the keyword index file 16 can be easily created.

FIG. 6 shows the relationship between records in the keyword table of contents file 16c.

In the keyword table of contents file 16c, the size of one record is fixed. At the beginning of each of the plurality of records, an item number (item number) uniquely determined for each record is added. One record corresponds to one keyword in principle. Therefore, since the record has a fixed length (that is, only a predetermined number of commands can be registered), when the number of commands exceeds a predetermined value, the command is truncated and not saved, and the corresponding command is not displayed even if searched. There is a possibility. Therefore, when the keyword index file 16c is registered with the number of commands for a keyword exceeding the limit value (predetermined number) of one record, the link data with the next record is also managed as the item number.
When the terminal 1 specifies a keyword, not only the command name associated with the keyword but also the command name associated with another keyword linked by the link data is specified. Therefore, even if the number of commands for each keyword increases, the display screen can be displayed in correspondence with an arbitrary number of commands without being truncated. The above link can be made between records as long as the file capacity permits.

In the uppermost record (first record) in FIG. 6, the item of the “link data of the next record”, that is, the item number is the item number of the second record (second record) from the top in FIG. Point. Thus, the keyword of the first record can store a command corresponding to the second record. Note that the keyword, item number, and item number of the second record are “Nu11”. This is because the second record follows the first record. However, the first
If a command needs to be further stored for the keyword of the record, the item number of the second record is made to indicate the item number of the record to be linked. Similarly,
You can link as many records as you need.

Also, as shown in FIG. 6, at the end of each record,
Matching link data corresponding to the keyword or item number is stored as an item number. This matching link data is
A keyword having some relation with the keyword is stored by its item number (item number). That is, if there is a keyword similar to the specified keyword, the file must be searched for the number of similar keywords, and there is a possibility that the command name cannot be quickly specified. Therefore, the keyword table of contents file 16 manages together the matching link data indicating the matching between the keywords by the item numbers. When the terminal 1 specifies a keyword, not only the command name associated with the keyword but also the command name associated with another keyword linked with the matching link data is specified. Therefore, when a certain keyword is specified in the terminal 1,
At the same time, another related keyword is specified, and the command name can be specified accordingly.

In the first record of FIG. 6, the item of the "matching link data", that is, the item number is the third record (the third record) and the fourth record (the fourth record) from the top in FIG. Indicates the item number. Therefore, each record is
A plurality of item numbers can be stored as the item number. The third record and the fourth record are different from the second record and are ordinary independent records. If the first record is specified by the keyword, not only the commands in the first record, but also the linked third and fourth
The command in the record can be specified at the same time. At this time, the command of the second record, which is a record following the first record, is also specified.

FIG. 7 shows an example of the matching link data. Now, the item numbers are from “001” to “006” for convenience. Since the keyword of item number “001” has no similar keyword, the matching link data is “Nu11”. No. `` 00
The keyword “2” has “003” and “005” as matching link data. The keyword of item number “003” has “006” (and “002”) as the matching link data. The keyword of item number “005” is “00” as the matching link data.
2 ”, ... The keyword of item number “006” has “003”,... As matching link data. As described above, when the keyword of the item number "002" is specified, the commands of the item numbers "003", "005", "006", ... are specified in addition to the command of the item number "002".

The command may be specified in consideration of the “depth of association” so that an infinite chain does not occur due to the matching link data. For example, when the keyword of the item number “002” is specified, only the item numbers “003” and “005” indicated by the direct match link data may be specified, or the item numbers “003” and “ The item numbers “006”,... Directly shown from “005” may be specified.

The present invention is characterized in that the link file 32 is provided on the center system 3 side as described with reference to FIG. As described with reference to FIG. 2, the link file 32 includes three files 32a, 32b, and 32c. The files 32a, 32b and 32c correspond to the files 16a, 16b and 16c in the terminal 1, respectively. Therefore, files 32a, 32b and 32c are each
The structure is the same as the structure of 16a, 16b and 16c.

As described above, in the center system 3, the keyword link index management file 32a and the keyword link index file 32b are provided in addition to the keyword link file 32c.
, It is possible to perform a high-speed search using a keyword.

Also, since the keyword link file 32c also manages link data, when a keyword is specified from the terminal 1 to the center system 3, one record is restricted not only by the terminal number and command name associated with the keyword but also by the link data. Other command names registered beyond the value can also be specified. When there is no corresponding command name in the terminal 1, the link data is received from the center system 3, and the terminal 1 receiving the link data receives the command name associated with the keyword of the linked item number or the keyword thereof. A desired command name can be obtained from the terminal 1 having the terminal number corresponding to the command number, and the characteristics can be obtained.

Further, since the keyword link file 32c also manages the matching link data, when a keyword is specified from the terminal to the center system, the link is made not only by the terminal number and command name associated with the keyword but also by the matching link data. The terminal number and the command name associated with another keyword can also be specified.
If there is no corresponding command name in the terminal 1,
The command name can be specified by the matching link data from the center system 3 or the desired command name can be specified from the terminal 1 by receiving the matching link data.

The link file 32 is similar to an integrated file of the table of contents files 16 of a plurality of terminals 1. Therefore, the center system 3
Is expected to be quite large. Further, it is expected that the search of the link file 32 in the center system 3 will take a considerable time.

Therefore, the structure of the keyword link file 32c is described in the eighth section.
A may be as shown in FIG. FIG. 8A is a comparison with FIG. 4, and shows the basic concept of the configuration of the keyword link file 32c. The keyword link file 32c is configured to manage the terminal number (TN) of the terminal 1 instead of the command name as information that can specify the terminal 1 in which each keyword is registered in units of keywords. The actual keyword link file 32c is
The structure is the same as that of the keyword table of contents file 16c shown in FIG. 5, except that a terminal number is stored instead of a command. Since the terminal number is shorter than the command,
The file can be reduced accordingly.

Conversely, when there is room in the file capacity of the center system 3, the configuration of the keyword link file 32c may be as shown in FIG. 8B. FIG. 8B is also comparable to FIG. The keyword link file 32c is configured to manage a command name associated with a keyword in addition to a terminal number.
FIG. 9 shows the relationship between the records in the keyword link file 32c in this case. FIG. 9 corresponds to FIG. The center system 3 notifies the terminal 1 of similar search to the terminal number of the terminal 1 that manages the command corresponding to the keyword, and this terminal number needs to be extracted from the normal command. Therefore, when a command is obtained as a result of the search, this extraction processing is required.
However, according to the file configuration shown in FIG. 8B, by obtaining the terminal number as the search result, the extraction process is eliminated, and the search is performed at a high speed.

By the way, the center system 3 includes the link file 32, and conversely, the terminal 1 including the link file 32 can be said to be the temporary center, that is, the center system 3. From this point, the center system 3 is also only one terminal, and is different from the other terminals 1 only in that a certain item of the distributed database, for example, a link file 32 for a display screen is provided.

An outline of the creation of the link file 32, that is, the setting of the center system 3 is shown in FIG. For example, terminal 1A
Declares to be a temporary center (center system 3),
This is notified to the other terminals 1B and 1C via the public network 4. In response to this, the terminals 1B and 1C respectively transmit the search data such as the table of contents and keywords relating to the data that may be shared (externally registered) to the terminal 1A via the public network 4 in their own databases. Send. The terminal 1A that has received the search information stores the search information in its own file. Here, the terminals 1B and 1C transmit the contents of their own component files 13 and only those of which the external registration is permitted (that is, the components whose flag can be set to “1”). Terminal 1A, based on the received information,
Create the link file 32. The terminal 1A also uses the contents of the component file 13 of the terminal 1A, for which the external registration is permitted, for creating the link file 32. Therefore, the link file 32 is obtained by integrating the parts of the table of contents file 16 of each of the terminals 1A to 1C with respect to the data permitted to be externally registered. The terminal 1A has a link file 32 in addition to its own table of contents file 16. By providing the link file 32, the terminal 1A is connected to the center system 3
Becomes

In addition, the terminal 1A creates the data table of contents 34 in the same manner as the link file 32. The data table of contents has, for example, a configuration substantially similar to that of the component file 13, and is a list of components (data) permitted to be externally registered. In FIG. 10, the data table of contents 34 shows only "screen name" and "version number" for convenience of illustration.

As described above, all the search information on the shared data in the database of each of the terminals 1A to 1C is registered in the center system 3 together with the corresponding terminal number (actually, the command), and is centrally managed. And a terminal, for example, a terminal
When a data search request occurs in 1B, first, the center system 3 is queried for search information, and the terminal 1C managing the desired data is searched for. Next, the terminal 1B can directly access the found terminal 1C and extract the desired data.

Since the terminals 1B and 1C other than the center system 3 do not manage the table of contents of the shared data of the entire distributed database, it is not possible to find the terminal number that manages the data required by each of the terminals 1B and 1C alone. Can not. However, according to the present invention, when the terminals 1B and 1C refer to the table of contents, they request the table of contents from the center system 3, receive the table of contents (part of the link file 32) from the center system 3, and By referring to the table of contents, it is possible to find the terminal number that manages the required data. When a keyword is used, the keyword is transmitted to the center system 3, and the center system 3 checks whether or not the keyword is applicable.
If "yes", the terminal number registered with this keyword is sent to the source terminal.

FIG. 11 shows communication between the center system 3 and the terminal 1.

The terminal 1A declares a temporary center and notifies the other terminals 1B and 1C that the terminal 1A has become a temporary center (center system 3).

The other terminals 1B and 1C that have received the notification transmit search information such as a keyword and a table of contents for the shared data in their own databases to the center system 3.

The center system 3 registers the received search information in its own link file 32 or the like. As described above, the search information regarding the shared data in the distributed database system is centrally managed by the center system 3.

The terminal 1B that has issued the data search request makes an inquiry to the center system 3 for search information.

Here, the center system 3 performs a process according to the content of the inquiry, that is, in the case of an inquiry using a keyword, searches this keyword from the registered data, and answers the corresponding terminal number to the inquiry source terminal 1B. If the keyword cannot be found, a response to that effect is made.

If the inquiry is not a keyword-based inquiry, the managed table of contents information is transmitted to the inquiry source terminal 1B. Terminal 1B
With reference to this table of contents information, a terminal that manages the required data is found.

The terminal 1B confirms the terminal 1C found in this way, requests data from the terminal 1C, receives the data, and outputs and displays the data.

As described above, search information such as a keyword and a table of contents for searching the shared database of the distributed database system is centrally managed by the center system 3 including the certain terminal 1. Therefore, each terminal other than the center system 3
It is sufficient for the terminals 1B and 1C to have a function of connecting and communicating with the other terminals 1C and 1B via the public network and the public network 4, and to retrieve data without imposing a load on the terminals 1B and 1C. It can be performed. Further, the terminals 1B and 1C other than the center system 3 may be constituted by a workstation and a personal computer alone.

The center system 3 has “version number” as information in the data table of contents 34 as shown in FIG.
The identity of data (parts) in each of the terminals 1A to 1C can be maintained. In other words, data can be created and updated only by the originating terminal. However, if data that is permitted to be externally registered (opened) is updated, data mismatch between the terminals 1A to 1C may occur. Can occur. In order to avoid this, the person who has updated the data must notify the terminals 1A to 1C of the change in the data, which is troublesome. Therefore, the "version number" is managed in the data table 34 of the center system 3 so that the terminals 1B and 1C searching for data can use the data table 34 to check whether the data is the latest data.

FIG. 12 shows a processing flow of data version number check.

A certain terminal 1B issues a command request to the center system 3 and acquires a command in order to refer to data created in another terminal 1C. Then, the database in the terminal 1B is searched by this command.
That is, it is checked whether or not the data exists in the file of the terminal 1B.

Since the origin of the data is the other terminal 1C, the own terminal
There is no such data in 1B. Therefore, terminal 1B
According to the processing in FIG. 11, the center terminal 3 creates the source terminal 1C.
And obtains the data from the terminal 1C (performs data learning). The learned data is displayed on the CRT of the terminal 1B in a later process.

If the data is data that has been previously learned by processing or the like, the data exists in the own terminal 1B.
Therefore, the terminal 1B refers to the "version number" of the data table 34 of the center system 3 to know the latest version number of the data.
Then, the version number of its own data and the center system 3
Compare with the version number obtained from.

Here, the version number of the data owned by the user can be known from the component file 13 for the data. Therefore,
This version number remains as it was when the data was obtained.
On the other hand, for the center system 3, the data creation source terminal
From 1C, a data change notification is issued each time data is changed. This notification is, for example, the data name (for example, screen name)
And the latest version number. Therefore, the version number of the data table of contents 34 of the center system 3 is always the latest. Therefore, if the data is changed after the terminal 1B learns the data, the version of the data owned by the terminal 1B differs from the version obtained from the center system 3 (the version of the data does not match). .

If the versions of the data match, the terminal 1B extracts the data from its own file. It is confirmed that this data is the latest data by matching the version numbers. This data will be used in the subsequent processing by the CRT of terminal 1B.
Will be displayed.

If the versions of the data do not match, the terminal 1B learns the creation source terminal 1C from the center system 3 and obtains the latest version of the data from the terminal 1C by the processing of FIG. 11 (relearning of data). I do). At this time, the learned data is displayed on the CRT of the terminal 1B in a later process.

The latest version of the data to be referenced is displayed on the CRT of the terminal 1B.

As described above, according to the present invention, even when data is changed, any of the terminals 1A to 1C can search and refer to the latest data. Further, the user does not need to be aware of the change of the data for referencing the latest data.

Next, a part registration process according to the present invention will be described with reference to FIG.

FIG. 13A shows an overall flowchart of the component registration process.

In a certain terminal 1, real data (for example, a screen) is created and registered as a component (step S1). This processing will be described later with reference to FIG. 13B.

Matching link data is set for the registered component (step S2). This process is shown in FIG. 13C.

It is checked whether or not the record of the keyword table of contents file 16c for the part exceeds a fixed record length of one record (step S3).

If it exceeds, the link data of the next record is set (step S4). This process is shown in FIG. 13D.
If not, step S4 is not executed.

The keywords are rearranged in a predetermined order, for example, in the order of the Japanese syllabary (step S5). This process is shown in FIG. 13E.

Among the registered parts, those for which external registration is permitted are designated and registered in the center system 3 (step S6). This process is shown in FIG. 13F.

FIG. 13B shows a part registration process, in which a registration screen is displayed and data is registered in the parts file 13 and the contents file 16.

When the actual data of the component is stored in the actual data file 20, the terminal 1 displays a registration screen for component registration on the display 10a (step S7).

Next, by interacting with the operator through the registration screen, the terminal 1 stores the component screen name,
The attribute information such as the comment, the command name (this is an instruction determined only by specifying what part), and the version number is registered (step S8). At this time, the command assignment unit 12 executes a command assignment process unique to the system. The version number is "1" because it is a new registration.

Subsequently, the terminal 1 registers one or a plurality of keywords assigned to the component to be registered in the keyword file 14 in association with the command name by the dialogue (step S9).

When the keyword file 14 is created in this way, the terminal 1 searches the keyword index index management file 16a shown in FIG. 5 to find out which item number of the first character each keyword in this file 14 is associated with. (Step S10).

Then, as shown in FIG.
However, each keyword is associated with a more detailed item number based on this item number and registered as a keyword index file 16b (S11).

From the keyword table of contents index file 16b and the keyword file 14 created in this way, as shown in FIG. 5, the table of contents file registration unit 17 stores the item number assigned to each keyword and the item assigned to the next keyword. The keyword number is registered and a keyword table of contents file 16c is created (step S12).

FIG. 13C shows the setting process of the matching link, selects the matched keyword, and displays the keyword index file.
16c shows the process of registration.

Keywords to be matched from the keyword index file 16b registered in step S4 (for example,
Elementary and junior high school students) and registered keywords (for example, students)
Is selected from the screen by the operator, and the item number is designated. As a result, as shown in FIGS. 6 and 7, the matching link data is added to the keyword table of contents file 16c, and the item number of the keyword to be matched is registered (step S13). This registration is performed by the table of contents file registration unit 17.

FIG. 13D shows a process of setting the record link data, and shows a process of registering in the next record when the number of commands corresponding to the keyword becomes too large to be registered in one record.

To add an item number when the number of commands exceeds one record, the table-of-contents file registration unit 17 expands the next record and, for example, generates an alarm to inform the operator that the number of commands exceeds one record. Notify (step S14).

Then, the table-of-contents file registration unit 17 registers the item number (item number) of the expanded record as the item number in the previous record (step S15).

If the number of commands is too large to accommodate the command in the expanded record, steps S14 and S15 are repeated until all the commands are registered.

FIG. 13E shows a file reordering process, and shows a reordering process of keywords newly registered in the keyword table of contents file 16c and the keyword table of contents index file 16b.

Also, the item number of the record in the order of the next Japanese syllabary is registered in each record of the keyword table of contents index file 16b (step S16).

Then, the item number of the next record in the order of the Japanese syllabary is registered as an item number in each record of the keyword table of contents file 16c (step S17).

By rearranging the records of the files 16b and 16c created in this way in the order of item numbers over a certain period (step S18), a keyword table of contents index file 16b and a keyword table of contents file 16c in alphabetical order are created. You. These processes are performed by the table of contents file registration unit 17.

FIG. 13F shows external registration processing, in which a part (screen name) for which external registration is permitted is specified, and a keyword and a command corresponding to this are transmitted to the center system 3 and registered.

When registering the data registered in the keyword index file 16c in the center system 3, the link file registration unit 15 displays a registration screen for external registration on the display 10a and interacts with the operator to register the data in the center system 3. A reception process for specifying a component, for example, a screen name is executed (step S19). As a result, the flag of the component file 13 corresponding to the screen name is set to “1”.

Next, the link file registration unit 15 searches for the keyword and command name assigned to the specified screen name by searching the component file 13 (step S2).
0).

Subsequently, the link file registration unit 15 adds the searched keyword and the command name (and the terminal number of the own station) and transmits the command to the center system 3 (step S21).

Then, when the center system 3 receives this transmission data, it checks whether or not the same keyword as the keyword requested for registration is registered in the keyword link file 32c. If registered, the command name of the terminal 1 requesting registration is registered at the entry of the keyword. That is, a command is additionally registered in an existing record. If not registered, the keyword and the command name for which registration is requested are registered in the keyword link file 32. As described above, the registration processing to the center system 3 is executed (step S22).

The reason why link data and item numbers are not transmitted to the center system 3 is that it is necessary for each terminal 1 to link only the parts whose external registration is desired in the center system 3. Therefore, thereafter, the center system 3 adds the link data in the same manner as the terminal based on the registered keyword link file 32c. In this case, the item number, the command, and the link data for each keyword may be registered in the keyword link file 32c. However, the terminal number is also registered in addition to the command, and the terminal number is directly transmitted from the center system 3. You may make it notify.

Then, the center system 3 can create the keyword link file 32c by registering the command with the terminal number added in the case shown in FIGS. 8B and 9.

FIG. 14 shows a component update process according to the present invention.

The actual data of the component is updated, and the updated actual data is registered in the actual data file 20 (step S23).

Next, the terminal 1 counts up the version number of the component in the component file 13 by +1 (step S2).
Four).

Thereafter, the registration processing described with reference to FIG. 13 is performed (step S25).

After the registration processing, the terminal 1
It is checked whether the flag of the part is “1” (step S26).

If the flag is "1", the component is a component that can be registered externally, and has been registered externally at the time of the previous registration.
Therefore, the terminal 1 notifies the center system 3 of the update. That is, a re-registration request is made to the center system 3 together with the screen name and the new version number (step S27). In the center system 3, the data table of contents 34 is updated to a new version number.

If the flag is "0", step S27 is not executed. This is because this component is not allowed to be registered externally and is data used only by the terminal 1 concerned.

Next, the fifteenth part search processing according to the present invention will be described.
This will be described with reference to the drawings.

 FIG. 15A shows the entire flow of the search process.

When searching and displaying a display screen of a desired component on the display 10a of the terminal 1, the terminal 1 calls up a keyword selection (designation) screen by interacting with the operator, displays a list of keywords, and displays a list of keywords to be searched. A keyword is selected (step S28).

For the selected keyword, terminal 1
The keyword index index file 16b of its own is searched to determine whether or not it exists in the file (step S29).

If it exists in the own terminal 1, a search process in the own terminal 1 is performed (step S30). This process is shown in FIG. 15B.

If it does not exist in its own terminal 1, it requests the center system 3 to perform a search, and the center system 3 performs a search process. Then, based on the search result from the center system 3, the terminal 1 requests data corresponding to the keyword from another terminal, and acquires the data (step S31). This process is shown in FIG. 15C.

The screen name of the component obtained in steps S30 and S31 is displayed on the display 10a (S32).

Usually, since there are a plurality of screen names displayed in step S32, the terminal 1 executes the selection of a command, that is, the narrowing down of the screen names by interacting with the operator as necessary (step S33). This process is shown in FIG. 15D.

The display screen specified by the narrowed screen name is displayed on the display 10a (step S34).

 FIG. 15B shows a search process in the own terminal 1.

By interacting with the operator, the terminal 1 calls the keyword table of contents index file 16b, displays a list of the keywords, and selects a keyword of the screen to be searched (step S35).

Next, the terminal 1 calls and searches the keyword table of contents file 16c, and specifies a command name (a plurality of command names) associated with the specified keyword (step S3).
6).

Next, the keyword index file in which the terminal 1 is displayed
It is checked whether or not an item number link between records exists on the screen of 16c (step S37).

If there is, the terminal 1 specifies the corresponding command name from the item number of the keyword indicated by the ring data (step S38).

That is, as shown in FIG. 5, for example, item number "0100020"
When the keyword “Ai” is designated, the terminal 1 also specifies the command of the next record order “0100021” according to the order of the keyword index file 16c.

On the other hand, when there is no link data, the terminal 1 skips step S38 and executes step S39.

Subsequently, the terminal 1 displays the displayed keyword index file.
It is checked whether a matching link exists in the screen of 16c.
(Step S39).

If there is, the terminal specifies the corresponding command name from the item number of the keyword indicated by the matching link data (step S40).

That is, as shown in FIG. 7, for example, when the keyword of the item number “002” is specified, the matching link data indicates the item numbers “003” and “005”.
This is equivalent to specifying three keywords of “2”, “003”, and “005”. Note that other keywords are registered in the matching link data, and similarly, the item numbers of the original keywords are registered in the matching link data of other keywords.

On the other hand, when there is no matching link data, the terminal 1 skips step S40 and executes skip S41.

The terminal 1 obtains the screen name associated with the command name specified in steps S36, S38 and S40 (step
S41), it is displayed on the display 10a (step S32).

Here, the screen name is specified from the command name for the following reason. As described above, the command describes the procedural meaning of the component. Therefore, the data processing unit in the terminal 1 and the center system 3
This command is sent and received as a message. Thereby, data processing can be performed. Therefore, the terminal 1 and the center system 3 basically handle components in accordance with the command, and when notifying the operator, display the equivalent of the command name in accordance with the screen name of the character string that the operator can understand. With this configuration, even when the screen name is changed later, the corresponding display screen can be specified without any error.

FIG. 15C shows a search process in the center system 3.

When the terminal 1 finds that there is no display screen for the component of the specified keyword in its own terminal, the terminal 1 executes a keyword specification receiving process by interacting with the operator, and transmits the received keyword to the center system 3. (Step S31).

The center system 3 searches the link file 32 in the same procedure as the search process in the terminal 1 shown in FIG. 15B, and specifies a command corresponding to the keyword (step S42). That is, the same processing as in step S35 is performed using the keyword link index file 32b, and steps S36 to S40 are performed using the keyword link file 32c.
The same processing is performed.

The center system 3 returns the command name specified in step S42 as a search result to the terminal 1 that has requested the search (step S43).

The terminal 1 having received the search result displays the received command name on the display 10a and selects one command name from a plurality of command names by interacting with the operator. Then, the command name is transmitted to the terminal 1 having the actual data of the component corresponding to the specified command to request data. (Step S44).

The terminal 1 that has received this command name reads the display screen data of the component associated with this command name from its own actual data file 20 via its own component realization table 21 and sends it to the requesting terminal 1. . When the keyword link file 32c of the center system 3 does not manage the command name, the terminal 1 reads the display screen of the component by asking the keyword to be transmitted.

Upon receiving the data, the request source terminal 1 stores the data in its own actual data file 20 and performs the registration processing shown in FIG. 13 on the data.

Thereafter, the terminal 1 obtains a screen name associated with the command name as in step S41, and displays the screen name on the display 19a (step S32).

As described above, since it is possible to know from the center system 3 which terminal 1 should execute the search process, the search of the display screen can be executed at a remarkably high speed.

In the case of a search for a plurality of display screens, the search request to the center system 3 is performed collectively.
Search results are also notified in a lump. Then, for example, the display of the received command name is repeated by the number of screens.

 FIG. 15D shows the command selection processing.

The screen names displayed in step 32 are usually plural. Therefore, the terminal 1 determines whether or not the operator makes a screen name narrowing request using another keyword (step S45).

When there is a request for narrowing down, the terminal 1 executes a receiving process for specifying the next keyword (step S46). Next, the terminal 1 searches the keyword contents file 16c to identify one or more command names associated with the specified keyword, obtains the corresponding screen name according to the component file 13, and obtains the corresponding screen name. The narrowing-down process is executed by specifying a screen name that is the same as the screen name that has been set (step S47). And terminal 1
Displays the narrowed screen name on the display 10a (step S48), and performs processing to return to step S24.

On the other hand, if there is no request for narrowing down in step S45, the terminal 1 determines whether there are a plurality of specified screen names (step S49), and whether there are a plurality of requests for display screens to be displayed. It is determined whether or not it is (step S50). When there are a plurality of specified screen names and a plurality of display screen display requests are specified, the terminal 1 opens a window and displays the specified display screen ( Step S51). On the other hand, when there are a plurality of specified screen names and there is a display request for one display screen, the terminal 1 executes the selection processing of one screen name from step 50 in accordance with the interactive processing with the operator (step 50).
S52), the display of the selected display screen is executed (step S53). And it is specified in step S49. When there is only one display screen, the terminal 1
Executes the display on the display screen (step
S53). Steps S51 and S53 overlap with step.

The specific display process of the display screen is as follows. The component realization table 21 is searched to specify the address information associated with the command name, and the display screen data of the corresponding component is retrieved from the real data file 20 according to the specified address information. Is executed by reading.

In this manner, according to the present invention, the corresponding command name can be specified only by accessing the keyword table of contents file 16 once, so that the display screen corresponding to the display 10a can be displayed.

[Industrial applicability]

As described above, the distributed database system and the data search processing method according to the present invention are useful when data search needs to be performed at high speed, and in particular, when a large number of data are connected via a public network. This is suitable for searching data released between terminals at a high speed while minimizing the burden on each terminal.

 ──────────────────────────────────────────────────続 き Continuing on the front page (31) Priority claim number Japanese Patent Application No. 1-339368 (32) Priority date Hei 1 (1989) December 27 (33) Priority claim country Japan (JP) (72) Inventor Former Keikei Kakehi 5-20-9-101, Edanami, Midori-ku, Yokohama-shi, Kanagawa Prefecture (72) Yoshio Mogi 102-4, Hirai-cho, Tochigi-shi, Tochigi (72) Takanori Fukatsu Miyazaki-665, Miyazaki-ku, Kawasaki-shi, Kanagawa-ken, Kanagawa-ken Megumi Mansion 101 (72) Inventor Masao Tomita 5-260, Mitsuhashi, Omiya City, Saitama Prefecture (72) Inventor Taku Oshiro 532 Shimoodanaka, Nakahara-ku, Kawasaki-shi, Kanagawa Prefecture Fujitsu Nakahara House W-8 210 Room (56) References JP-A-61-142587 (JP, A) JP-A-63-5440 (JP, A) JP-A-1-9521 (JP, A) JP-A-62-1198929 (JP, A) JP-A-62-243033 (JP, A) JP-A-63-142428 (JP, A) JP-A-62 125445 (JP, A)

Claims (13)

(57) [Claims] 1. A distributed database system wherein each of a plurality of terminals (1) has a distributed individual database and is connected to a public network (4), wherein each of said terminals (1) Is a keyword table of contents index management file and a keyword table of contents index that manages the command name to which the keyword is assigned in units of a keyword, which is one of the attribute information for the display screen and also one of the attribute information. Table of contents file (1
6) and the center system (3), which is one of the terminals (1),
It has a link file (32) for managing information specifying a terminal (1) in which the keyword is registered in units of keywords, and the keyword table of contents index management file is a file in which the first character of a keyword is different from the first character Is a file that manages a starting item number indicating the beginning of the keyword index file. The keyword table of contents index file is more detailed based on the item number of the first character registered in the keyword table of contents index management file. The keyword index file manages keywords and command names, and the keyword index file stores the item numbers registered in the keyword index index file, and which keyword is next to a certain keyword. Link information that manages whether to continue Distributed database system, characterized in that it. 2. The distributed database system according to claim 1, wherein said command name can function as information for specifying said terminal (1). 3. The link file (32) is composed of a keyword link index management file, a keyword link index file, and a keyword link file, and the keyword link index management file is a file in which the first character of a keyword is different from the first character. The keyword link index file is a file that manages a starting item number indicating the beginning when the keyword link index file is more detailed based on the item number of the first character registered in the keyword link index management file. The keyword link file manages a keyword and a command name, and stores the item number registered in the keyword link index file and a keyword next to a certain keyword. Distributed database system according to claim 1, wherein the de is a file that manages the link information for managing whether subsequent. 4. The keyword table of contents file (16c) comprises a plurality of records. The record can register a predetermined number of command names, and when the command name corresponding to the keyword exceeds the predetermined number, the keyword 2. The distributed database system according to claim 1, wherein the link data to the next record is registered in the record where the record exists. 5. The keyword contents file (16c) includes a plurality of records. When there is another keyword related to one keyword, the record of the one keyword contains the other keyword. 2. The distributed database system according to claim 1, wherein link data matching a record with the record is registered. 6. A command name assigned to each terminal (1) connected to the center system (3), the command name being assigned to data in a database and being an attribute of the data and specifying a terminal that manages the data; One or a plurality of keywords set in association with the command name are managed, and when a keyword is specified, the command name associated with the specified keyword is specified, so that the data of the corresponding data is specified. A data search processing method for a distributed database system for executing a search, wherein each of the terminals (1) includes a keyword table of contents (16c) that manages command names to which the keywords are assigned in units of keywords. A key for the center system (3) to manage a terminal number of a terminal (1) in which the keyword is registered in units of a keyword; A terminal file (32c), and when a keyword is specified, the terminal (1) specifies a command name associated with the keyword according to its own keyword contents file (16c), or Keyword link file (32c)
Searching the data by specifying a command name associated with the keyword according to the keyword table of contents (16c) possessed by the terminal (1) having the terminal number specified by (1). Data retrieval processing method of the system. 7. The keyword link file (32c)
Manages the command name associated with the keyword in addition to the terminal number. When the keyword is specified, the terminal (1) is associated with the keyword according to its own keyword index file (16c). The data search is performed by specifying a command name or by specifying a command name assigned to a terminal number specified by the keyword link file (32c). Data retrieval processing method for distributed database system. 8. The keyword table of contents (16c) further manages an item number paired with a keyword, and matching link data indicating matching between keywords by the item number. When a keyword is specified, not only the command name associated with the keyword but also the command name associated with another keyword linked with the matching link data is specified, and the item number paired with the keyword is temporarily determined. 7. The data search processing method for a distributed database system according to claim 6, wherein the keyword index file (16c) is created after preparing the keyword index index file (16b). 9. The keyword link file (32c)
When the matching link data is also managed and a keyword is designated from the terminal (1) to the center system (3), the link is linked not only with the terminal number and command name associated with the keyword but also with the matching link data. 9. The data search processing method for a distributed database system according to claim 8, wherein a terminal number and a command name associated with another keyword are also specified. 10. The keyword table of contents file (16c)
Further, when the number of commands corresponding to the keyword and the number of commands for the keyword exceed the limit of one record are registered, link data to the next record is managed together with the link data indicated by the item number, When the terminal (1) specifies a keyword, not only the command name associated with the keyword but also other command names linked by the link data are specified, and the first character of the keyword is associated with this. A keyword table of contents index management file (16a) that manages the assigned item numbers, and a keyword table of contents index file (16a) that manages keywords based on the item numbers of the management file (16a).
7. The distributed database system according to claim 6, wherein the keyword table of contents file (16c) rearranged in a predetermined character order by the management file (16a) and the table of contents index file (16b) is created. Data retrieval processing method. 11. The keyword link file (32c).
Manages link data together, and when a keyword is designated from the terminal (1) to the center system (3), the link is linked not only with the terminal number and command name associated with the keyword but also with the link data. 11. The data search processing method for a distributed database system according to claim 10, wherein the other command names are also specified. 12. The center system (3) is one of the terminals (1), and is a terminal (1) that has declared to be a temporary center. A data search processing method for a distributed database system according to any one of the preceding claims. 13. A distributed database system in which a plurality of terminals (1) each having a distributed individual database are connected to a public network (4), wherein each of the terminals (1) has a keyword unit. A terminal (1) having a keyword table of contents (16c) for managing the command names to which the keywords are assigned, and in which the keywords are registered in units of keywords;
Keyword link file (32
c), wherein one of the terminals (1) declares to be a temporary center (3) and transmits the terminal number of the temporary center (3) to another terminal (1); ) Transmits search information for searching data which may be shared in its own database to the temporary center (3), and the temporary center (3) registers the received search information, The terminal (1) that needs the data is the temporary center (3)
Using the search information to search for a terminal (1) in which necessary data is managed, accessing the searched terminal (1), and searching for the data. Data retrieval processing method.