JPH02238568A - Data base processing system - Google Patents

Abstract

PURPOSE:To improve the input/output processing efficiency and to reduce the load on a central processor by receiving a page which is edited by an auxiliary storage controller by indicating the processing of a data base after the central processor supplies column information and search information to the auxiliary storage controller. CONSTITUTION:The auxiliary storage controller 2 reads a data base specified by the central processor 2 and stores it in its internal storage means, columns are classified from respective records of the stored data base according to column information, and then columns specified by the search information are selected. Data which meet requirements specified by the search information are extracted from the selected columns and edited into pages, which are sent out to the central processor 1, so only data that the central processor 1 requires in the data base read out of the auxiliary storage device 3 are sent out. Consequently, the amount of data transferred between the auxiliary storage controller 2 and central processor 1 decreases to improve the input/output processing efficiency, the central processor 1 requires neither retrieval nor extraction of necessary data from the data base, and the load is reduced.

Description

[Detailed description of the invention] [Table of contents] Overview Industrial Icheon field Prior art Problems to be solved by the invention Means for solving the problems Actions Examples Effects of the invention [Summary] Stored in auxiliary storage device To improve input/output processing efficiency and reduce the burden on the central processing unit, regarding a database processing method that can reduce the burden on the central processing unit of a computer system that reads and processes a database, and also increases input/output processing efficiency. With the aim of Column information without format,
A computer system equipped with a central processing unit that processes a database read from an auxiliary storage device using comparison data, comparison conditions, and search information indicating a logical condition that is a combination of each comparison condition. When instructed to process the database, stores the database read from the auxiliary storage device in the internal storage means, classifies the columns in each record of the database read into the storage means based on the given column information, Based on the given search information, an auxiliary storage control device is provided that selects records that meet the specified conditions from the selected column and edits them into a page, and the central processing unit transfers the column information and search information to the auxiliary storage control device. The configuration is such that the auxiliary storage control device receives the edited page by instructing the database processing.

[Industrial application field]

The present invention relates to a disk storage processing method that can reduce the burden on a central processing unit of a computer system that reads and processes disk storage stored in an auxiliary storage device such as a disk device, and can improve input/output processing efficiency.

With the development of the information society, the amount of data processed by computer systems has increased, and various software has been developed for processing this database. The database is read into the main memory, and then the central processing unit searches and extracts the necessary data, so it takes time to transfer the data from the auxiliary storage to the main memory. Moreover, the processing time of the central processing unit is also becoming longer.

Therefore, it is desired to reduce the burden on the central processing unit and to improve input/output processing efficiency.

[Conventional technology]

FIG. 4 is a circle illustrating an example of a database.

A database is made up of multiple pages, each page being a collection of multiple records, and each record being a collection of multiple columns. Therefore, a column is the minimum data unit of the database, and the columns of each record contain data such as name, age, address, etc., respectively.

Such a database is recorded, for example, in a disk device connected to a computer system. Therefore, the central processing unit reads the database from the disk device, stores it in the main memory, and then performs a search according to instructions from the program to extract the desired data.

That is, using column information indicating column length and format,
For example, if the purpose is to obtain demographic statistics by dividing columns in a record containing consecutive numbers and analyzing whether the divided columns are numbers or letters, etc. , search the column that is the age data of each record, and extract the population by age in a certain area using the comparison data of the search information, the comparison conditions, and the logical conditions that are a combination of each comparison condition. Or, perform processing such as extracting the population of a certain age or older in a certain area.

[Problem to be solved by the invention]

As mentioned above, conventionally, after reading the database from the disk device to the main storage, the central processing unit searches for and extracts the necessary data from the main storage, which eliminates unnecessary data. In other words, records that include unnecessary columns that are not extracted or unnecessary records that do not correspond to the data required by all columns are transferred to the main storage, increasing the amount of transferred data and causing problems for people. At the same time as the output processing efficiency decreases, the main memory 2 performs the search and extraction process for data required by the central processing unit.
There is a problem in that the load on the central processing unit is heavy.

In view of these problems, the present invention provides column information and search information from a central processing unit to an auxiliary storage control unit such as a disk control unit, so that the auxiliary storage control unit can perform the tasks required by the central processing unit. To reduce the amount of data to be transferred by editing and sending data, and to eliminate the need for search and extraction processing in a central processing unit, thereby increasing human output processing efficiency and reducing the burden on the central processing unit. It is an object.

[Means to solve the problem]

FIG. 1 is a bronch diagram of the principle of the present invention.

The central processing unit 1 is connected to an auxiliary storage device 3, which stores a database as shown in FIG. 4, through an auxiliary storage control device 2.

Then, the column information and search information are sent to the auxiliary storage control device 2. After sending it to the control device 2, an instruction is given to perform database processing.

When the auxiliary storage control device 2 is instructed by the central processing unit 1 to perform data storage processing, the auxiliary storage control device 2 reads the designated database from the auxiliary storage device 3 and stores it in internal storage means. Then, based on the given column information, the columns in each record of the database read into the internal storage means are classified, and based on the given search information, one '-search information is extracted from each record of the database. Select the column you want to specify, select the records that match the conditions specified in the search information from the selected column, and edit them on the page.

Then, when the data has been edited for a predetermined number of pages or when the editing process is finished, the auxiliary storage control device 2
Send out page by page sequentially.

[Effect]

By configuring as described above, the auxiliary storage control device 2
reads out and stores the database specified by the central processing unit 1 in an internal storage means, classifies columns according to column information from each record of the stored database, and selects the column specified by the search information; Data matching the conditions specified by the search information is extracted from the selected columns, edited into pages, and sent to the central processing unit 1 from the database read from the auxiliary storage device 3. Only data required by the central processing unit 1 is sent.

Therefore, the amount of data transferred between the auxiliary storage control device 2 and the central processing unit 1 is reduced, the human output processing efficiency is improved, and the central processing unit 1 can search and extract necessary data from the database. It is not necessary and the burden is reduced.

(Embodiment) FIG. 2 is a block diagram of a circuit showing an embodiment of the present invention.

The second circle shows an example of a disk control device including a plurality of functional units each having a processor and having independent functions. The channel adapter 5 includes an interface control circuit 11, a data transfer control circuit 12, and a common bus control circuit 13, which are transferred by the processor 10.
is provided, and the commands from the channel 4 of the host device inputted via the interface control circuit 11 are sent to the processor l.
For example, if the storage device O receives an instruction to transfer data to and from the disk device 7, the disk drive adapter 6 that connects the specified disk device M7 via the common bus 31 via the common bus control circuit 13 Send the address of

The high-speed adapter 6 has the same configuration as the channel adapter 5, and when the processor 14 receives an address from the channel adapter 5 via the common path control circuit 15 of the high-speed adapter 6, it is sent via the common bus control circuit 15. It is coupled with the common nozzle control circuit 13 of the channel adapter 5. Then, it controls the disk device 7 via the interface control circuit 17, and notifies the processor 10 of the channel adapter 5 when it becomes possible to write or read data at the specified address.

Here, the block sensor 10 of the channel adapter 5 controls the data transfer control circuit 12, and the processor 14 of the high speed adapter 6 controls the data transfer control circuit 16, so that when an instruction to write data is received from the channel 4, In this case, the channel adapter 5 requests data transmission from the channel 4l, and sends the data transmitted from the channel 4 to the interface control circuit 11, the data transfer control circuit 12, the common bus control circuits 13 and 15, the data transfer control circuit 16, and the interface. The data is transferred to the disk device 7 via the control circuit 17 and written.

Also, if data reading is instructed, the data read from the disk device 7 is transferred in the opposite direction to the above,
Send to channel 4.

When a central processing unit (not shown) performs database processing, it notifies the processor 10 of the channel adapter 5 via the channel 4 that column information is to be sent.

Processor 10 of channel adapter 5 that received this notification
sends an address to and connects to the input/output port 19 of the shared storage 8 via the common lotus control circuit 13 and the common lotus 3I, and writes the column information sent by the channel 4 to the huffer memory 18.

That is, the processor 1/10 of the channel adapter 5 transfers column information sent from the channel 4 to the common lotus 31 via the interface control circuit 11, data transfer control circuit 12, and common lotus control circuit 13 to the buffer memory via the input/output port 19. Write to the 180 column information group area.

Next, channel 4 is connected to Prosensor 10 of channel adapter 5.
After that, to the processor 14 of the high speed adapter 6,
Instructs to read the database from the disk device 7.

Therefore, if the high speed adapter 6's IJ14 is
Through the common bus control circuit 15, the input/output baud 1-]96 address of the shared storage 1/rage 8 is sent and combined, and the specified decous is successively sent from the disk device 7, and the input/output baud 1-]9 Hashifa Memory 18 via 19
Write to the input page group area of .

Channel 4 then sends out search information and instructs the start of database processing.

Therefore, the processor 10 of the channel adapter 5 is connected to the processor 2 through the common bus control circuit 13, through the common bus 31 and the input/output ports I/29 of the database assists 1 and 9.
Send the address to 3, combine it, and database assist l
- Write the search information sent by the channel 4 into the search information storage area of the Huffer memory 30 of 9.

At this time, the fluoro sensor 10 of the channel sensor 5 is
Based on instructions from channel 4, database assistance 1
- Instruct the processor 23 of No. 9 to read the column information written in the column information group area of the Hashfer memory 18 of the shared storage 8.

Therefore, the processor 23 sends the address from the input/output boat 24 to the input/output boat 20 of the shared storage 1/rage 8, connects it, accesses the column information group area of the buffer memory I8, reads out the column information, and The column information is written to the column information storage area of the memory 25.

When the processor 14 of the device adapter 6 completes writing the data for J pages shown in the page of FIG.
The completion of data writing for one page is notified to the Brosenser 23 via the common bus control circuit 15, the common bus 31, and the input/output port 24 of the database assist 9.

Therefore, the processor 23 accesses the puffer memory 18 via the input/output ports 1 and 24 and the human output ports 20 of the shared storage 1 and the storage 8, reads data from the input page group area, and passes through the input/output ports 24 to the buffer memory 25. The data is written to the input page storage area of .

When the data of 1.1 yen is written to the input page storage area of the Hanofa memory 25, the analysis processing unit 27 of the database assist 1-9 sequentially imports records from this data and analyzes the column groups that make up the record. The record of consecutive numerical values is broken down into column φ positions by comparing the column appearance stored in the column information storage area of the Hanwha memory 25 with the column information consisting of the column number and format, and it is determined whether the broken down column is a number or a character. Perform an analysis to determine the format of etc.

Then, this analysis result is sent to the page creation processing section 28.

In addition, for the selection processing unit 26, the Huffer memory 25 of all the columns read out from the Huffer memory 25 and analyzed
Sends the start address stored in , the column number specified in column information, and the length of the column.

This process is repeated until all records of the input page stored in the Huffer memory 25 have been analyzed.

Regarding column length, the method for specifying column length differs depending on the structure of the column.
Some are specified in the column information, some are specified in the fixed area of the column itself, and some have their length determined regardless of the length specified in the column information or the column itself. The analysis processing unit 27 determines the length of each column and breaks down records containing consecutive numerical values into columns.

The selection processing unit 26 selects the search information stored in the search information storage area of the Hanwha memory 30 from among the selection information composed of the start address, column number, and column length sequentially sent by the analysis processing unit 27. Extracts those that match the column number specified by .

Then, using this extracted selection information, comparison data obtained from the search information, comparison conditions, and logical conditions that are a combination of the respective comparison conditions, one page of data stored in the input page storage area of the Huffer memory 25 is extracted. The power rams are taken in sequentially from 1 to 3 and the comparison process is executed. In other words, the column length of the specified column number is recognized from the selection information, and a comparison process is performed to determine whether the record to be selected is, for example, a record in which a column in which a specified age in a specified region is recorded. Determine whether or not.

The logical conditions for this comparison process are made up of logical expressions of several comparison results. When the comparison process is completed, for example, when the comparison process for the address column is completed and the comparison process for the age column is not yet completed, it can be determined whether the record being processed is a selection target or a non-target record. That is, it is determined whether a record is to be selected or not, depending on whether the column contents included in the record satisfy the comparison condition.

Selection processing unit 2 6 1 # Once it is determined whether the processed record is a selection target or not, the analysis processing unit 27 and page creation processing unit 2B are notified of the result.

The analysis processing unit 27 and the page creation processing unit 28 will either interrupt the processing if the record being processed or has completed processing is not to be selected by the notification from the selection processing unit 26, or change the processing to the record that has been completed. is invalid and the next record is not read from the Hanwha memory 25. Further, if the item is selected and is being processed, the process continues, and if the process is completed, the page creation processing unit 28 performs the following process.

The page creation processing unit 28 performs processing to create a new basis in the output page storage area of the puffer memory 30. That is, only the columns to be extracted sent by the analysis processing unit 27 are sequentially stored in the Huffer memory 30, and the analysis processing unit 2
When receiving a notification from 7 that processing for one record has been completed, and receiving a notification from the selection processing unit 26 that a processed record is to be selected, it is determined that one record for output has been completed, and this record is In order to store the record in the huffer memory 30, it writes to the area specified by the first storage address, allocates an area for storing a new record, and moves on to the next record creation process.

Such processing continues until no data is stored in the input page storage area of the Huffer memory 25, and the amount of records stored in the output page storage area of the Huffer memory 30 is monitored by the processor 23.

When one page of records is stored in the output page storage instruction page area of the Huffer memory 30, the Brosensor 23 connects the input/output port 29 to the shared storage 8 via the common lotus 31.
Huffer memory 30
The records read from the buffer memory 18 are written to the output page group area of the buffer memory 18.

The processor 10 of the channel adapter 5 is the processor 10 of the channel adapter 5.
From 3, 1 to 8 buffer memories for shared storage 1 to storage 8
Either the user is notified that a page's worth of data has been written, and the notifications are accumulated to reach a predetermined number of bases, or the output page storage area of the buffer memory 30 becomes empty, and database assistance}-9 When it is notified that the processing of Then, multiple pages are transferred to channel 4 in succession.

FIG. 3 is a time chart illustrating the example of the operation shown in FIG.

If the vertical axis represents processing and the horizontal axis represents time, then process (2) represents the time required for device adapter 6 to read the database from disk device 7 and write each page to the input page group area of buffer memory 18 of shared storage 8. show. In other words, a case is shown in which pages 0 (from input page 1 to [phase]) are sequentially stored.

Processing @ is performed by the processor 23 of the database assist 9.
It shows the time required for each page from the input page group area of the Huffer memory 18 to enter the input page storage area of the Huffer memory 25. That is, the data of input pages 2 to [phase] is transferred after a delay of one page from process 2.

Processing ■ indicates the time required for column comparison processing by the analysis processing section 27 and the selection processing section 26 of the database assist 9.

That is, it shows the time taken by the selection processing section 26 to determine whether or not the selection target is selected and to notify the analysis processing section 27 and the page creation processing section 28 of the result. The gap between the manual pages ■ and ■ indicates that the comparison process up to the input page ■ is completed before the processing of the input page ■ starts.

Processing [phase] is performed by the page creation processing unit 28 using BasofaMemo IJ.
Store records sequentially on 30, input page ■ ~ [phase]
This shows the time it takes to edit 10 pages of 1 to 3 output pages ■゛~■゜.

Processing [phase] is executed by the processor 2 when one page worth of output pages is completed in the output page storage area of the Hashfer memory 30.
3 indicates the time to transfer the output page to the output page group area of the huffer memory 18 of the shared storage 8.

Processing [phase] is carried out by the processor 1o of the channel adapter 5,
This shows the time required to read three output pages from the output page group area of the buffer memory 18 of shared storage 1-rage 8 and transfer them to channel 4.

As described above, according to this embodiment, the time required for data transfer due to data processing is reduced to 3/10.

Since the processes @ to (2) are executed in parallel with the data transfer shown in process (2), the processing time is apparently almost zero.

〔Effect of the invention〕

As explained above, the present invention reduces the amount of data transferred between the channel and the disk controller, so it is possible to improve input/output processing efficiency, and the central processing unit can collect necessary data. Since it is sent from the disk controller, the load can be reduced.

[Brief explanation of drawings]

Fig. 1 is a block diagram of the principle of the present invention, Fig. 2 is a circuit diagram showing an embodiment of the present invention, Fig. 3 is a time chart explaining the operation of Fig. 2, and Fig. 4 is a block diagram of the principle of the present invention.
The figure is a diagram illustrating an example of a database. In the figure, 2 1 to 1 are central processing units, 2 are auxiliary storage controllers, 3 are auxiliary storage devices, 4 are channels, 4 is a channel,
5 is a channel adapter, 6 is a storage adapter, 7 is a disk device, 8 is a shared storage, 9 is a database assist 1-, 10, 14.23 is a processor, II, 17 is an interface control 1 output circuit, 12. 16
is a data transfer control circuit; 13. 15 is a common lotus control circuit; 18, 25. 30 is eight sofa memory, 19-22,
2 units, 29 is the input/output board 1, 26 is the selection processing unit,
27 is an analysis processing section, and 28 is a page creation processing section.

Claims (1)

[Claims] An auxiliary storage device (3) that stores records made up of a collection of columns of minimum data units, pages made up of the collection of records, and a database made up of the collection of pages. and read from the auxiliary storage device (3) using column information indicating the length and format of the column, and search information indicating the comparison data, comparison conditions, and logical conditions that are a combination of the respective comparison conditions. In a computer system equipped with a central processing unit (1) that processes the database, when an instruction is given from the central processing unit (1) to process the database, the auxiliary storage device (3) is stored in the internal storage means.
), classifies the columns in each record of the database read into the storage means based on the given column information, and searches the database from each record based on the given search information. The central processing unit ( After 1) gives the column information and search information to the auxiliary storage control device (2), it instructs database processing so that the auxiliary storage control device (2) receives the edited page. A database processing method characterized by: