JP3316763B2 - Storage controller - Google Patents

Description

Description: BACKGROUND OF THE INVENTION The present invention relates to a storage device, and more particularly, to a storage control device for controlling a disk device.

[Conventional technology]

The disk device is used as a storage device of a computer system. The disk device is generally managed by being divided into cylinder, track, and sector areas. For example, data stored in a corresponding area can be accessed by designating one sector of m tracks of n cylinders.

The cylinders, tracks, and sectors described above indicate physical areas of the disk device, and the user manages data for each file. The file includes a large number of record data required by the user and the like, and includes, for example, a sales totaling file, a sales file for each person in charge, an item file, and the like.

When writing or reading a file, a system such as an OS manages which sector of which track of which cylinder in the disk device writes the file. The sector is generally one data block of 128 bytes, 256 bytes, 1024 bytes or more, and a plurality of sectors exist in one track. However, not all the data of the aforementioned file is stored in one sector, but is stored in a plurality of sectors.

[Problems to be solved by the invention]

The files managed by the system such as the OS described above are stored in a plurality of sectors. However, the files are not always written continuously to the continuous sectors of the track in the disk device, and are stored discontinuously over a plurality of tracks. Is done. In particular, when a file stored up to now is deleted, the area where the file is stored becomes an open area, and when a new file is written next, the file is first written to the open area, and then When the area is not enough, the remaining data is stored by writing it to a sector in another open area. Therefore, when writing and deletion of a plurality of files are repeated, the sectors of each written file are not written continuously, and must be accessed by moving the head, for example, in units of one sector. For this reason, there has been a problem that the access speed is reduced and the disk management by the OS or the like becomes complicated.

Also, the file stored in the disk device is composed of a plurality of records. If the records are arranged in the order of the record key and stored, the management of each record in the file is simplified, and the record obtained by specifying the record keyword is determined. Can be accessed. The problem at this time is when a record is deleted or a new record is added. In this case, the record group must be rearranged in the order of the record keywords each time.
It takes an enormous amount of time for disk processing for this purpose. In particular,
When each record is composed of variable-length data, each record can be stored in the order of the record keyword without any gaps, so the storage efficiency is high, but not only when records are deleted or added, but also when If the record length changes as a result of correcting or totalizing, all the record groups must be rearranged in the order of the record keywords.

An object of the present invention is to make it possible to read out target data from a disk device at high speed, and to store data without lowering storage efficiency.

[Means for solving the problem]

The storage control device according to the present invention is configured such that a plurality of records before update and input information including update type information for updating the plurality of records are used as an update record in a storage area different from the plurality of records before update as a track unit. Based on the position information of the pre-updated record stored in the disk device and the position information indicating the storage area of each of the updated records, based on the position information of the position storage means Reading the pre-update record and the update record from the disk device in track units, comparing the read record keys of a plurality of pre-update records and the update record in each read track, and finding the same record key as the update record. Update that updates the record before the update according to the update type information of the update record And storing the updated record updated by the record updating means and the pre-update record not updated by the record updating means in the storage area of the disk device following the storage area where the updated record is stored in track units. When the updated record and the non-updated record are written to the storage area in track units by the updating means, the individual storage position information of the record position storage means is replaced with position information indicating a new storage area. Rewriting position information management means.

[Function block diagram]

 FIG. 1A is a functional block diagram of the present invention.

The disk device 1 is a storage device that stores pre-update record data in the order of record keys.

The data position storage means 2 stores storage position information of the pre-update record data in the disk device 1.

The update record storage means 3 stores an update record including update type information for updating the contents of the record data before update.

The updating means 4 reads the record data and the updated record from the disk device 1 and the updated record storage means in track units, updates the record before the update according to the update type of the updated record, and sets the updated record in the order of the record key. The data is again written to a new storage location of the disk device 1.

The location information rewriting means 5 rewrites the storage location information of the data location storage means 2 with the new storage location information written by the updating means 4 in the previous term.

FIG. 1B is a second functional block diagram of the present invention.

The disk device 6 is a storage device that stores record data.

The disk control means 7 controls the reading and writing of data on the disk device 6 in track units. The memory 8 stores, for example, data for at least one track in writing and reading.

The processing means 9 performs processing based on the data for one track, and stores the memory 8 in the disk control means 7 upon completion of the processing.
Of the disk device 6 in the memory 8 is added.

〔Example〕

Hereinafter, an embodiment will be described with reference to FIGS. 2 to 8. FIG. 2 is an overall circuit configuration diagram. The data processing circuit 10 is a circuit including, for example, a microprocessor or the like, and performs various processes. The data processing circuit 10 includes a work RAM 11, a read data RAMA12, and a read data RAMB1.
3. The disk control circuit 14 and the write data RAM 15 are connected.

When the data processing circuit 10 is performing some kind of processing, for example, totaling sales, a temporary calculation result of each totaling is stored in the work RAM 1.
Stored in 1. That is, the data processing circuit 10 performs processing using the work RAM 11 as a work area. Then, when the obtained result is data to be stored in the disk 16, the result is stored in the write data RAM 15. When data to be processed is stored in the disk 16, the data processing circuit 10 reads the data for one track in which the target data is stored through the disk control circuit 14 to the disk control circuit 14. , Read data RAMA12 and read data RAMB13
To be stored. The data processing circuit 10 performs processing using the data for one track stored in the read data RAMA12 and the read data RAMB13.

Prior to storing data consisting of a plurality of words of one record on the disk, the data is stored in the write data RAM 15. In this case, if the next one word cannot be stored in the write data RAM 15, the data processing circuit 10 Stores a track end code after the last data in the write data RAM 15. And 1 in the write data RAM 15
A control signal is input to the disk control circuit 14 in order to store the data for the track in the track following the track on which the data has been written on the disk 16. This disk control circuit
When the control signal is applied to the disk control circuit 14, the disk control circuit 14
Controls writing of data for one track in the write data RAM 15 to the disk 16.

The disk 16 stores a large number of files composed of a plurality of records, and the data structure of each file is configured as follows. That is, if the file delimiter code is represented by “§”, the record delimiter code is represented by “(”, and the word delimiter code is represented by “,”, the file delimiter code “§” is added to the beginning of each file, followed by the file The name is
Next, the first keyword RK ₁ following the record delimiter code “(”, followed by the word delimiter code “,”, the second keyword RK ₂ , and the data words separated by the following word delimiter code “,” When a plurality of records exist in one file, a record delimiter code "(" is placed at the beginning of each record, and the first and second keywords are similarly separated by a word delimiter code. Further, a plurality of data words are separated by a word separating code, wherein each of the separating codes is represented by a specific code which is not used for data used for each word.

The files are successively stored on the disk 16 in order of file number (file name), and further, in the same file, each record is stored in ascending or descending order of the keyword. As described above, the processing of storing in the order of keywords is performed by the data processing circuit 10.

Each file is the second and subsequent tracks on disk 16
are stored in t ₁ , t ₂ ,. In this storage, the track t ₀ of the beginning of the disk track number indicating the position of the file number and the file of a file containing is stored is stored.

Figure 3 is a data configuration diagram of a track t ₀ of the beginning. Track number t ₁ of the beginning of the files described above represented by t _F for this track t ₀ is the start track number t ₁₇ maintenance record group in turn represented by t _M is shown followed by t _E track number t ₂₀ the end of the maintenance records group, further, then the file number §f
Track numbers t ₁ , t ₇ , t ₁₁ storing ₁ , §f ₂ …
... is set. As the head track t next track t ₁ after the _zero are shown in Figure 6, the track t ₁ ~t ₁₆ is stored in the file group, the track t ₁₇ ~t ₂₀ is stored in the group maintenance records I have. For example the hard disk controller 14 as the file number and the corresponding track number of files that need to their aggregation is read from the beginning of the track t ₀ accesses the track number when the data processing circuit 10 performs processing of summary etc. Operate. Then, the target data is stored in the read data RAMA12 or the read data RAMB13 for each track, and is processed.

In the present embodiment, all file accesses are performed in track units, and writing and reading of files are performed at a higher speed than in sector units.

On the other hand, it is necessary to perform maintenance such as correcting the written data. Generally, a record to be modified is read and modified, and the modified record is written in the same area again.In the embodiment, the modified record is used as a maintenance record.
The data is written to tracks other than the area where the data has been written. Therefore, the maintenance record is managed by the file number to be corrected and the keyword, and the correction data is handled in record units.

FIG. 4 is a flowchart for writing a maintenance record. The data processing circuit 10, when a request such as correcting a record stored in the disk 16, totaling data in a record, deleting a record, or adding a new record occurs, generates a Perform the processing shown. A new maintenance record having information to be corrected is input from a keyboard (not shown) and stored in the read data RAMB13 as shown in FIG. That is, the file separation code “§”, the file number “f ₂ ”, the record separation code “(”, and the keywords RK and R
K ₂ , update type n, and data words m, ‥‥‥ are written. The maintenance record is provided with a file number, a keyword, and an update type, and this record indicates which record of which file is to be updated and how. The update type n is information indicating any of addition, correction, deletion, and aggregation. Adding this maintenance record to the end of the maintenance record group t ₁₇ ~t ₂₀ in the disc shown in FIG. 6 with the following operation.

First read the track number t ₂₀ represented by the track number t _E at the end of the group maintenance records from the track t _0, further stored in the read data RAMA (RA) 12 reads the data in the track of the track number t ₂₀ ( S1 in the flow of FIG. 4). Furthermore, the read data RAMA (R
A) Store the contents of 12 in the write data RAM 15. By this processing (S1, S2), the maintenance record in the last track of the maintenance record group is stored in the write data RAM (W
M) Stored in 15. Then, whether the remaining area exists in the write data RAM 15, that is, the data of the first word of the maintenance record in the read data RAM 13 Ҥ
It is determined whether or not there is an area in which f ₂ , ”and the track end code can be written (S3). When the remaining capacity exists, the data of the subsequent maintenance record can be written. First, one word of data is read from the beginning of the maintenance record in the read data RAMB13 (S4). It is stored at the position following the word (S8). Then, it is determined whether all words of the maintenance record have been completed (S9). That is, it is determined whether or not the remaining word exists in the read data RAMB13. If all the words have not been completed (N), the processing (S3 to S9) is performed again.

On the other hand, when it is not within the remaining capacity in the determination (S3) (N)
Cannot write the word that follows it, so it will be written on the next track. Therefore, first, the track end code Is written to the position following the last word in the write data RAM 15 (S5), and the process of writing the data in the write data RAM 15 to the track indicated by the track number t _E is performed (S5).
6). In the above-described processing (S6), the data processing circuit 10 applies a write control signal to the disk control circuit 14, and the control signal causes the disk control circuit 14 to write data RA.
The contents of the M15 performs a series of processing for writing the track t _E in the disk 16. End track number after processing S6
After performing the process (S7) of incrementing the track indicated by t _E by 1 (S7), the processes (S3 to S9) are performed again. When writing of all the maintenance records in the read data RAMB13 is completed by the above-described processing (S9), the write data RAM
Storing a file delimiter code "§" on the area following within (WM) 15 and (S10), followed by writing the maintenance record in the write data RAM15 to track indicated by the track number t _E. Then, the maintenance record writing process ends (END). In the process (S10), the process of storing the “§” code in the work RAM 11 is performed to indicate that the maintenance record is the last by the file separation code “§”. Is a null code, for example, 0, it is determined to be the last.

FIG. 6 is a diagram showing the data structure written to each of the tracks t _{1 to} t ₂₀ in the disk device 16. Starting track t ₀
The track index and file index as shown in Figure 3 is stored, without gaps each record consisting of variable length data to the keyword order for the truck t ₁ ~t ₁₆ in the track i.e. FIG. 6 following the track This is a group of files in which the packed and stored files are written in file number order. The track of the track number t ₁₇ ~t ₂₀ is a track for storing group maintenance records. The t _M and t _E shows the track number at the end of the start track number and maintenance record group of group maintenance record in the third diagram the above-described case of FIG. 6 the track number t ₁₇ and the track number t ₂₀ is stored.

The maintenance record group is added to the next track of the file group by the processing described in the flow of FIG.

FIG. 7 is a flowchart of a file group reading process. This processing is performed when the data processing device 10 needs a group of files stored in the disk 16 in advance in processing such as tabulation. First, to store the contents of a track of the track number t ₀ the read data RAMB (RB) 13 (S1
2). Subsequently, a file number in the read data RAMB (RB) 13 corresponding to the file number of the processing target specified in advance is detected, and a track number corresponding to this file number is searched from the data in the read data RAMB 13 (S13). The contents of the track indicated by the track number are read from the disk 16 and read data RAMA12 is read.
(S14). By this processing (S14), the data of the first track of the target file is read data RA.
Stored in MA12. Then, read / write data RAM1
Reads the records in 2 in units of one record
Are stored sequentially (S15). At this time, the record Code (ie, track end code) is determined (S16), and If the code does not exist (none), it is determined whether the next file number exists (S17). Then, when the next file number does not exist (none), the processing (S1
Repeat from 5). On the other hand, in the determination (S16) When it is determined that the code exists (present), the read data RA
Since the reading of the data in MA12 has been completed and the file to be read exists following the next track,
After the next track number update process (S18), the updated track is read. That is, a process of storing the track data indicated by the track number from the disk 16 to the read data RAMA (RA) 12 is performed. By this determination, a file over a continuous track can also be read.

On the other hand, in the determination (S17), when it is determined that the next file number exists (Yes), it means that all the records of the file to be processed have been read out to the work RAM 11, so that the totals of this record, for example, Processing such as tabulation of records having the same keyword is performed (S19), and all processing ends (END).

As described above, in the embodiment, data is handled in file units. Since these files are successively stored in the order of the file numbers in the track, access is possible in track units, and high-speed reading can be performed. Further, since records are sequentially stored in the file in the order of keywords, the process of reading records of specific keywords is also performed at high speed. The above-described processes such as file aggregation are performed after the file update process using the maintenance record group is completed.

FIG. 8 is a flowchart of a process for updating a maintenance record. When starting the update process, first, track number is set corresponding to the track number i.e. t _M and t _E shows the beginning and end of the maintenance record group stored in the track number t _0, in the example of FIG. 6 t
₁₇ ~t reads ₂₀ records stored in tracks, sorted in file number order More Description order, rewrites the track from t ₁₇ of the disk 16 to t ₂₀ (S21). In this case, when the changed track number of the end of group maintenance records, writes the track number in association with t _E to track t _0. Then, lead the contents of the track of the track number t ₀ data
After storing in RAMA (RA) 12, t _F indicating the beginning of the file group
In the register F, the track number corresponding to t _M indicating the beginning of the maintenance record group in the registers M and M ′, and the track number corresponding to t _E indicating the end of the maintenance record group in the register E. Is stored in the register E '(S22). This process (S2
2) is an initial setting process to be used for a process described later.

All files updated in the update process are written from the track following the track at the end t _E of the maintenance record group. The next process S23 is a process for obtaining a track in which the updated file is newly stored,
The track number of the register E is incremented. Then, the data stored in the track indicated by the track number in the register F is read out and stored in the read data RAMA12 (S24). That is, since the first track number of the file group is stored in the register F, the data in the first track of the file group is read out by the processing S24 and stored in the read data RAMA12.
Since the first track number of the maintenance record group is stored in the register M, in the process (S25), the data in the first track of the maintenance record group is stored in the read data RAMB13.

In this way, the read data RAMA12, read data
The RAMB 13 stores the data of the track at the beginning of the file group and the data of the track at the beginning of the maintenance record group. Using this, first, matching processing of each record is performed in the order of keywords, and the result is stored in the write data RAM 15 (S26). In this matching processing, a maintenance record of the same keyword as the record in the read data RAMA12 exists. When there is no record, the record in the read data RAMA12 is stored in the write data RAM 15 as it is, and when it exists, the process according to the update type of the maintenance record (S27) is performed. For example, if the update type is tally, the corresponding word of the maintenance record is added to the corresponding word of the file record, and if the update type is correct, the corresponding word of the file record is corrected to data of the corresponding word of the maintenance record, If it is added, the maintenance record is added as a file record, and if it is deleted, the record of the file having the same keyword as the maintenance record is deleted. Thereafter, it is determined whether the file number of the processed record is the same as the file number of the previously processed record (S28). This determination (S28) is for determining whether processing of a new file has been performed or whether the processing is a continuation of the previous processing. When the files are not the same file (N) in this determination (S28), the file number and the track number in the register E are stored in the register MM in association with each other as a file index. When the same file is determined in the determination (S28) (Y), or after the process (S29) is completed, it is determined whether the read data RAMA12 is empty, that is, whether the matching process has been performed for all the records in the read data RAMA12. It is determined (S30).

In the process S27, each word of the updated record is sequentially written to the write data RAM 15.At this time, if the next word to be written and the track end code cannot be written to the remaining area of the write data RAM 15, the following S
After executing the processes of 31 to S34, the remaining words are again written to the write data RAM 15 in the process S27, and after the writing of the updated one record is completed, the read word in the read data RAMA12 or the read data RAMB13 to be processed is written. Erase records. If it is determined in the determination (S31) that data cannot be written in the write data RAM 15 (N), the data is written after the last word data in the write data RAM 15 After the code is stored (S32), control for writing the contents of the write data RAM 15 to the track specified by the register E is performed (S33), and the process of incrementing the contents of the register E (S34) is performed. Processing (S2
Perform 7). In the determination (S31), the write data RAM1
When the updated record in 5 has been completely written, the matching process (S26) and subsequent processes are performed. In the determination (S30), when the read data RAMA12 is empty (Y), that is, when all the file data of one track that has been read have been updated, the process of incrementing the track number in the register F (S35) is performed. The read data RAMA12 stores the file data for one track before the update. After the update, the contents of the register F are incremented to read the file data of the next track. Then, the contents of the register M 'are compared with the contents of the register F "(S36). The register M' stores the beginning track number of the maintenance record group before updating by the processing (S22) (start of the maintenance record group). When the content of the register F is equal to or larger than the content of the register M ', it indicates that all the reading of the file group has been completed. When the contents of the register F are not greater than or equal to the value of the register M '(N), that is, when the contents are smaller than the value of the register M', the contents of one track indicated by the track number of the register F are read, and the read data is read. Stored in RAMA12 (S3
7) Then, the discrimination (S38) is performed. The determination (S38) determines whether or not the read data RAMB13 is empty. When it is not empty (N), that is, when there is still a maintenance record to be processed, the processing after determination (S31) is performed. On the other hand, when the maintenance record does not exist and is empty (Y), the register M is incremented (S39).
Then, it is determined whether the result is larger than the value of the register E '(S40). The register E 'stores the end track number of the maintenance record group before updating in the processing (S22), and when the content of the register M is not larger than the content of the register E' in the discrimination (S40) (N), When the size is smaller, the remaining maintenance record group exists on the disk 16, so the processing (S25) is performed again, the maintenance record for one track of the disk 16 is stored in the read data RAMB13, and the above-described processing is repeated.

When the content of the register M is larger than the value of the register E '(Y), it means that the processing such as addition or deletion of all the maintenance records has been completed, and then the same determination (S41) as the determination (S36) I do. When F ≧ M ′ is not satisfied, the processing following the processing (S31) is executed again.

A new updated record is stored in the track of the disk 16 indicated by the register E in the order of the file number and in the order of the keyword by repeating the above-described processing.
In addition, in processing S34, S35, S39, each register E,
F, track number in M is when it becomes a value that exceeds the last track number of the disk 16 is to be reset to the track of t _1.

When the content of the register F is greater than or equal to the content of the register M 'in the determination (S41) (Y), that is, an update process such as rewriting the record designated by the maintenance record is performed, and all the files are updated. Is completed, the file division “§” indicating the end is stored in the write data RAM 15 (S4
2) The contents of the write data RAM 15 are stored in the track specified by the contents of the register E (S43). And processing (S4
After 3), create a track index (S
Perform 44). As shown in FIG. 3, the track index is composed of the first track number of the file group, the first track number of the maintenance record group, and the last track number of the maintenance record group. In this embodiment, as described above, the updated new file group is stored from the track next to the track number at the end of the maintenance record group before updating. Therefore, track index creation processing S4
In 4, the end track number of the maintenance record group before the update is stored (performed in step S22).
This is a value obtained by incrementing the content of the register E '. That is, the track number t _F is a value obtained by adding 1 to the track number in the register E '. The updated last track number is stored in the register E, and the value obtained by adding 1 to the value is the start track number t _{M of} the maintenance record group and the end track number of the maintenance record group.
t _E. That is, in a state where no maintenance records stored in this state, after the maintenance record written according to the flow of FIG. 4 mentioned above is written from the track indicated by t _E. Storing data of the file indexes created into the process at track index and processed was created (S29) In this indexing process (S44) is the track t ₀ (S4
5) Yes. After the above-described processing, the data processing circuit ends a series of update processing of the maintenance record.

As a result of this update processing, the maintenance file becomes 0 and becomes a completely new file. Such processing is performed by, for example, performing total processing, managing sales inventory, and the like in one file, assigning a keyword to each record, and performing the update processing at the end of the day, for example, to update the file. Can be. In addition, since only data to be updated, which is generated in one day's work, is created and registered as a maintenance record, and is updated collectively later, the processing speed is much faster than changing each time. Things.

In the above-described embodiment of the present invention, whether data in one track can be written or not is determined in word units. However, the present invention is not limited to this, and may be managed on a record basis. In the embodiment, the records are stored in the order of the file numbers and in the order of the keywords. However, the present invention is not limited to this. Any record may be stored at least in the same file in the order of the keywords.

In the above-described embodiment, the storage position of the new file after the maintenance record is updated is controlled so as to be stored following the track next to the last track in which the maintenance record was stored. First, the file storage area of the disk is divided into approximately equal halves, and the positions t _F and t _F ′ of the respective leading tracks are stored as indices. and the track position of t _F at the beginning, the following stores files on each track sequentially stores the position t _M of the next track of the last track of the file maintenance records sequentially. Then, the updating of the maintenance record, the new file after updating sequentially written to the beginning of the track position of the t _F '. The maintenance record entered after this update is the position of the track following the last track of this file, t _M ′.
From the beginning. t _F a new file of the result of updating the stored maintenance records from the position of the 'stored files and t _M from the position of' write sequentially from the position of the t _F. Thereafter, each time the file is updated, the updated file is written from the position of t _F ′ or the position of t _F alternately. This makes it easier to manage the storage locations of the files on the disk.

〔The invention's effect〕

According to the present invention, when updating a plurality of pre-update records stored in the disk device, the pre-update record and the update record stored in a storage area different from the storage area of the pre-update record are track-by-track units. Only the records to be updated with the same read and record keys can be updated according to the update type included in the update record, and the updated record after the update is stored in the storage area of the update record Since the data is stored in the storage area in units of tracks, there is no need to issue an update processing command to update the records, and all the pre-update records except for the updated records stored in the disk device need to be rewritten. And the records are updated in the order of the record keys, so that the updating process can be performed in a short time.

[Brief description of the drawings]

1 (a) and 1 (b) are functional block diagrams of the first invention and the second invention, respectively, FIG. 2 is a circuit configuration diagram, FIG. 3 is a data configuration diagram of a first track, and FIG. FIG. 5 is a block diagram of a maintenance record, FIG. 6 is a block diagram of a disk track, FIG. 7 is a flowchart of a file read process, and FIG. 8 is a flowchart of a maintenance record update process. 1 ... disk device, 2 ... disk control means, 3 ... memory, 4 ... processing device.

Continuation of the front page (56) References JP-A-53-79445 (JP, A) JP-A-52-79916 (JP, A) JP-A-52-50130 (JP, A) JP-A-63-36375 (JP, A) , A) JP-A-62-31426 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06F 3/06, 12/00

Claims (1)

(57) [Claims] 1. A plurality of records before updating and input information including update type information for updating the plurality of records are stored as update records in a storage area different from the plurality of records before updating on a track basis. A disk device, position storage means for storing position information indicating a storage area of each of the record before update and the update record stored in the disk device, and the disk based on the position information of the position storage means. The record before the update and the update record are read from the device in track units, and the read record keys of the plurality of pre-update records and the update records in each read track are compared. Record update means for performing an update process on a record in accordance with the update type information of the update record; Update means for writing, on a track basis, a post-update record updated by a record update means and a pre-update record not updated by the record update means in a storage area of the disk device following the storage area where the update record is stored. Position information management for rewriting individual storage position information of the record position storage means to position information indicating a new storage area when an updated record and an unupdated record are written to the storage area in track units by the update means; Means, and a storage controller.