JP3515581B2 - Data search method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [0001] BACKGROUND OF THE INVENTION The present invention relates to an external storage device.
Information processing to search for data stored in
Data retrieval method in a physical system. [0002] 2. Description of the Related Art FIG. 7 shows an example of a conventional information processing system.
FIG. 8 is a diagram showing the sequential search file in FIG.
FIG. 9 is a diagram showing an example of a file structure, and FIG.
FIG. 10 is a diagram illustrating an example of a search process. FIG.
FIG. 11 is a diagram showing an example of a hexadecimal search file structure, and FIG.
FIG. 12 is a diagram showing an example of a binary search process in FIG.
8 shows an example of a hash search file structure in FIG.
FIG. 13 is an example of the hash search process in FIG.
FIG. In FIG. 7, an information processing system 10 comprises:
Processing device (CPU) 1, storage device (MM) 2, magnetic disk
Disk unit (DK) 3, magnetic disk controller (DKC)
5, input / output device (IO) 6 and input / output control device (IO)
C) 7 and the processing device (CPU) 1
Data search request input from the input / output device (IO) 6
In the magnetic disk drive (DK) 3
Retrieve the specified data from file 4 and, for example,
In response to a data registration request input from the input / output device (IO) 6
Accordingly, the designated data is registered in the file 4. In a conventional information processing system 10,
Is a sequential search process as a data search process in the file 4.
, Binary search processing and hash search processing
Was. First, an example of a conventional sequential search process is shown in FIGS.
And FIG. [0005] File 4 shown in FIG.
A (Each data is A_i, Where i indicates the order of registration)
It is arranged in the order of recording, and data A₁Is the fastest climb
It is recorded data and data A_nWas last registered
Data. At the beginning of file 4, the current file
The number n of registered data is stored in the file 4. Take
In the state, any data A from the input / output device (IO) 6_Xof
When a search request is input, the processing device (CPU) 1
Data A registered in file 4_iExtract and enter in the order of registration
Search target data A_X(Step S9 in FIG. 9)
1 to S93), search target data A_XData that matches
A_iThe extraction and comparison are repeated until
S95), matching data A_iIs detected (step
S93), detected data A_iRequired information corresponding to
Report D_i(Step S96), and all data A_i
Does not match (step S94), the search target data
A_XIs not registered (step S97). [0007] New data A_YRegister to file 4
The data stored at the beginning of file 4
By referring to the data number n, the last data A_nIs registered
The n-th area that has been registered is identified, and the registration target data A_Y
In the next (n + 1) th area_{n + 1}Stored as
At the same time, the number of data n is updated to (n + 1). As is clear from the above description, the conventional order
In the next search process, new data A_YRegistration process
Always refers to the number of data n and the data A_YRegistration area
Recognizes and stores data A in the recognized registration area._YTo store
Although the registration time is constant and short, data A_XInspection
The search processing is performed on the search target data A._XData A that matches_iBut
Search time to search sequentially from the first area until detected
The number of accesses to file 4 that determines
A_XAnd the average number of access
It becomes longer in proportion to the data number n. The processing device (CPU) 1 has a file 4
Extract a predetermined number of data from a single access
Or store, but the number of accesses is proportional to the number n of data
The tendency is not to change. Next, an example of a conventional binary search process is shown in FIG.
This will be described with reference to FIG. 10 and FIG. Files shown in Figure 10
4, each data B (individual data B_jWhere j is ascending
Order) are classified into a certain order (ascending or descending).
For example, assume that data B is sorted in ascending order.
Then, data B₁Is the currently registered minimum data B,
Data B_nIs the currently registered maximum data B. At the beginning of file 4, the current file
The number n of registered data is stored in the file 4. Take
In the state, any data B from the input / output device (IO) 6_Xof
When a search request is input, the processing device (CPU) 1 first
Referring to the number of data n, all data B registered in file 4
Intermediate value B of_m(For example, when the number of data n = 7, the data B
_Four, And when the number of data n = 8, the data B_FourEtc.)
And search target data B_X(Step S11 in FIG. 11)
1 and S112), search target data B_XIs the intermediate value B_m
If smaller, intermediate value B_mLower area bisected by
Intermediate value B _mExtract and search target data
B_X(Steps S113 and S112).
Search target data B_XIs the intermediate value B_mMedium if larger
Intermediate value B_mIs further medium for the upper half area divided by
Intermediate value B_mTo extract the search target data B_XAnd compare
S114 and S112), search target data B_XWhen
Matching intermediate value B_mRepeat the above process until is extracted
And search target data B_XIntermediate value B that matches_mIs extracted
(Step S112), the extracted intermediate value B_mTo
Required required information D_i(Step S11)
5). Further, new data B_YRegister to file 4
If you want to register
Data B_YDetermine the registration area of the
All the data B stored in the
The registration area determined by moving
Data B_YIs stored in the registration area, and the number of data n is
Update to (n + 1). As is apparent from the above description, the conventional two
In binary search processing, data B_XSearch processing of the registration
Search time while searching for all data B
The number of accesses to file 4 to be determined is also in the order of registration and ascending
It is almost constant regardless of the order, and increases directly with the number of data n.
Relatively average search time without any proportional increase
, But new data B_YRegistration process
After the registration area is determined in the same process as the
Move all data B stored in the destination to one data destination
Required, the longer the registration area is near the top, the longer the registration time
I do. Next, an example of a conventional hash search process
This will be described with reference to FIGS. 12 and 13. Shown in Figure 12
File 4 has a hash value search area A_hAnd the data area
Area A_chAnd the chain data area A_chxAnd is provided
You. Hash value search area A_hIs the search target
Data C (each data is_j) Is predetermined
By assigning to the hash function, the corresponding hash
There is an area for calculating the value h, for example, each search target data C
_jAnd the calculated hash value h_jAnd an index table corresponding to
Is considered. Data area A_chCorresponds to each data C
Area stored in the area specified by the hash value h
And the hash value h_jData corresponding to_hjAnd table
You. Chain data area A_chxIs the data distribution
Multiple data, such as when the number of data is large relative to the length of the column
Is likely to be the same
Thus, the same hash value h was obtained from a plurality of data C.
In this case, a plurality of data C corresponding to the hash value h are linked.
Data area A_chWithin this
Seed hash value h_jData C corresponding to_hj1, C
_hj2,... A pointer P indicating the storage area of the concatenation of_chjIs stored
Have been. In this state, the input / output device (IO) 6
Arbitrary data C_XWhen a search request is input, the processing unit
(CPU) 1 first of all, hash value search area A of file 4
_hTo retrieve the search target data C_XHash value h of_XSeeking
(Step S131 in FIG. 13), then the data area A_chof,
Hash value h_XRefers to the area specified by
Step S132), the only data C in the area_hXIs stored
If there is (step S133), the extracted data
C_hXRequired information D corresponding to_X(Step S
136). On the other hand, the data area A_chThe hash value h_X
Pointer P to the area specified by_chXIs stored
If there is (step S133),
Area A_chxAnd pointer P_chXSpecified by
Data C stored in the area_hXExtract one to
Search target data C_X(Step S134
And S135), matching data C_hXIs detected
And the extracted data C _hXRequired information D corresponding to_XSee
(Step S136). Further, new data C_YRegister to file 4
If you want to register
Data C_YHash value h corresponding to_YAnd the data area
Area A _chHash value h in_YIn the area corresponding to
Data C_YIs stored. Note that the hash value h_YBut other registration
Data C_hjIf they match, the data area A_chTo
Pointer P_chAnd the registration target data C_YIs Choi
Data area A_chxTo be stored. As is clear from the above description, the conventional c
In the hash search process, the data C_XThe search process of
Search time to determine the search area by finding the hash value h
The number of accesses to file 4 that determines
Although search time is reduced, c
When the hash value is the same, the number of searches increases,
The time may be long. Also new data C_YClimb
Recording has the same problem. [0021] The problem to be solved by the present invention is apparent from the above description.
As described above, in a conventional information processing system,
When using a search, the registration time for new data is uniformly short
Search time is longer depending on the number of data and the order of registration.
Fluctuate, and when using a binary search, the search time
Is almost uniform in a short time, but the registration time of new data is large or small.
When it fluctuates greatly in order and further uses hash search
Has a long search time and registration time depending on the data structure.
Search time and registration time are both short
Time was not enough. According to the present invention, regardless of the number and configuration of data,
The aim is to reduce search and registration time
You. [0023] FIG. 1 shows the principle of the present invention.
FIG. In FIG. 1, reference numeral 100 denotes an object of the present invention.
Information processing system, 200 is an information processing system 100
Is a file to be searched which is provided in. Reference numeral 210 denotes a file in the file 200 according to the present invention.
Are a plurality of data areas provided. 220 is
According to the present invention, each data area 210 is stored in the file 200.
This is a spare area provided correspondingly. Reference numeral 201 denotes a file in the file 200 according to the present invention.
Common to each data area 210 and spare area 220
This is the first data area that has been shifted. 300 according to the invention
Search means provided in the information processing system 100.
You. Reference numeral 400 denotes an information processing system according to the present invention.
100 is a registration means. 500 is the present invention
Re-editing means provided in information processing system 100
It is. [0027] Each data area 210 has a predetermined number of data.
Data storage capacity, and multiple
Store the data respectively. The spare area 220 is the same as the data area 210.
It has a single data storage capacity and the storage area of the data area 210
When the data area becomes insufficient, the
Storage area 220, data area 210 and spare area 22
The data is classified and stored in a predetermined order from 0 through 0. The first data area 201 is composed of each data area 2
At the top of the array of data stored in
Only the first data to be stored is stored and arranged in
You. The search means 300 determines that a search request for required data is information
When input to the processing system 100, the first data area
Searches the area 201 and retrieves the data in which the search target data is stored.
Data area 210 and the spare area 220 are detected.
Search from the issued data area 210 or spare area 220
Search for the target data. The registration means 400 is used to register required data.
RequestSaidWhen input to the information processing system 100,
Searches the first data area 201 and stores the registration target data
Search for a pair of data area 210 and spare area 220 to be
Data area 210 or spare area 220
Are stored in descending order. The re-editing means 500 is provided for any data area 2
Additional data is registered in the pair of 10 and spare area 220
When there is no room left
Spare area 220 of a pair of area 210 and spare area 220
The data stored in the new data area 210 and
And moved to the data area 210 of the pair of the spare area 220
Later, a pair of a new data area 210 and a spare area 220
All data areas 210 and spare areas 220 including
Re-edit the data stored in the pair in descending order, and
Start of new data area 210 in start data area 201
Add data and sort by size. The data area 210 and the spare area 22
0 indicates that the information processing system 100
And the number of data that can be stored and extracted in one access
Determining the storage capacity so that it can be stored is considered. The search means 300, registration means 400 and
And the reediting means 500, each data area 210, the spare area
When searching for the first data area 220 and the first data area 201,
Searching by a binary search method is considered. Therefore, the data search range is a pair of data
Search time is significantly longer because it is limited to the area and spare area
It is shortened and free space for registering new data
Since the number of data to be moved for generation is also limited, register
The time will be greatly reduced, and the information processing system will be
System data retrieval and registration efficiency is greatly improved. [0035] An embodiment of the present invention will be described below with reference to the drawings.
You. FIG. 2 shows an information processing system according to an embodiment of the present invention.
FIG. 3 is an example of a file structure in FIG.
FIG. 4 is a diagram showing the data search process in FIG.
FIG. 5 is a diagram showing an example, and FIG. 5 shows a new data registration in FIG.
FIG. 6 is a diagram showing an example of processing, and FIG.
FIG. 14 is a diagram illustrating an example of a file reediting process. In addition, through all figures
The same reference numerals indicate the same objects. FIG. 2 shows the information processing system shown in FIG.
Information processing system shown in FIG.
An information processing system 10 having the same configuration as that of FIG.
FIG. 3 shows a file 200 shown in FIG.
File 40 in the magnetic disk drive (DK) 3
Provided, and as a search means 300 in FIG.
Search unit 11 for executing a data search process as shown in FIG.
Is provided in the processing unit (CPU) 1 and is also shown in FIG.
The new data as shown in FIG.
The registration unit 12 that executes the registration process is a processing device (CPU) 1
And the re-editing means 500 in FIG.
To execute a file re-editing process as shown in FIG.
The reediting unit 13 is provided in the processing device (CPU) 1
You. The file 40 shown in FIG.
K basic data areas having the specified data storage capacity N
Area A_B(Individual basic data area is A_BfWhere f
Are provided with 1 to k), and each basic data area A_BTo
Correspondingly, the basic data area A _BSame data storage capacity as
Spare area A with N_SIs provided for each one, and each basic
Data area A_BAnd spare area A_SCommon to the first data
Area A_HAnd final data storage area A_TEach one
Are provided. Each basic data area A_BAnd spare area
A_SThe data storage capacity N of the processing device (CPU) 1 is
File40Can be stored and extracted once access to
It is assumed that the number is set equal to the number of usable data. All data E stored in the file 40
Are sorted in ascending order and then divided into k or less data groups G.
And the basic data area A_BIs stored in the
Area A_BIf it cannot be stored in the
Area A_STo be stored. Therefore, each data group G₁, G_Two, ..., G_k
Any data extracted from is always arranged in ascending order
You. Each data E included in each data group G is
Are sorted in ascending order, and the corresponding basic data areas A
_BAnd, if necessary, spare area A_SIs also stored. What
Basic data area A in data group G_BPart stored in
The basic data group G _BAnd reserved area A_SStored in
Part is the spare data group G_SCalled. Each basic data area A_BEach basic data
Data group G_BNumber of data contained in_BIs stored in each spare
Area A_SContains each preliminary data group G_SNumber of data included in
n_SIs stored. On the other hand, head data area A_HContains the basic data
Area A_BfBasic data group G stored in_Bfin
, And each head data E sorted at the top in ascending order_BfH(Ie
Each basic data group G_BIs the smallest data in the
They are arranged in ascending order. Further, the final data storage area A_TIn the file
When all the data E stored in the file 40 are arranged in ascending order,
The final data E that is finally classified_T(Ie the largest de
Data) is stored. First, a file search process according to the present invention
This will be described with reference to FIGS. 2 to 4
And any data E from the input / output device (IO) 6_XSearch for
When a request is input, the processing device (CPU) 1
Start 1 The started search unit 11 first stores the file 4
0 leading data area A_HTo condition E_BfH≦ E
_X<E_{B (f + 1) H}Data E where_BfHSearch for
As a result, the search target data E_XMay be stored
Basic data area A_BfAnd spare area A_{Science fiction}Decide
(Step S41 in FIG. 4). Next, the search unit 11 determines the determined preliminary area A_{Science fiction}
And the stored spare data group G_{Science fiction}At the beginning
Data E_SfHAnd extract the search target data E_XAnd size comparison
(Step S42). As a result of the comparison, the condition E_X≧ E_SfHHolds
For example, search target data E_XIs the spare area A_{Science fiction}Stored in
(Step S43), the preliminary data group G_{Science fiction}From
Search target data E by binary search_XExtract (Step
S44), extracted search target data E_XRequired for
Information D_XIs referred to (step S46). As a result of the comparison, the condition E_X<E_SfHIs established
Then, search target data E_XIs the basic data area A_{Science fiction}Case
Is determined (step S43), and the basic data
Area A_BfTo access the basic data group G_BfExtract and extract
Released basic data group G_BfSearch target data by binary search from
Data E_XIs extracted (step S45), and the extracted search pair
Elephant data E_XRequired information D corresponding to_XRefer to
Step S46). As is clear from the above description, the present embodiment
According to the data E_XThe search unit 11 for searching for
First, the head data area A of the file 40_HAccess
Search target data E_XBasic data area A in which is stored
_BfAnd spare area A_{Science fiction}And then reserve area A_{Science fiction}To
Access and search target data E_XIs the basic data area A _Bf
And spare area A_{Science fiction}Judge which is stored in
And spare area A_{Science fiction}Is determined to be stored in
Is immediately the spare data group G_{Science fiction}Search target data E from _XExtract
Out, basic data area A_BfIs determined to be stored in
Data area A_BfAccess to basic data
Group G_BfSearch target data E from_XTo extract
The number of accesses to the file 40 may be two or three times,
Search time is also reduced. Next, a new data registration process according to the present invention will be described.
This will be described with reference to FIGS. 2, 3 and 5. Figures 2, 3 and
In FIG. 5, new data is input from the input / output device (IO) 6.
TA E_YWhen a registration request is input, the processing device (CPU)
1 activates the registration unit 12. The started registration unit 12 first stores the file 4
0 final data storage area A_TAssessed and stored
Last data E_TAnd the registration target data E_YAnd large
A small comparison is made (step S51 in FIG. 5). As a result of the comparison, the condition E_Y<E_TWhere is established
In this case (step S52), the registration unit 12
Area A_HTo access each head data E_BfHExtract
And condition E_B1H> E_YIf the condition is satisfied,
TA E_YBasic data area A for storing_B1And determine the article
Case E_BfH≦ E_Y<E_{B (f + 1) H}Data E where_Bf
HIs searched, the registration target data E_YStore
Basic data area A_BfAnd spare area A_{Science fiction}Determine (S
Step S53). Next, the registration unit 12 determines the reserved area A_{Science fiction}
And the stored spare data group G_{Science fiction}At the beginning
Data E_SfHTo extract the search target data E_XAnd size comparison
And, as a result of the comparison, the condition E_Y> E_SfHIs satisfied,
S54), search target data E_XIs the spare area A_{Science fiction}Case
Is determined to be stored, and the spare area A_{Science fiction}Number of data extracted from
n_{Science fiction}And preliminary data group G_{Science fiction}To be registered by binary search
Data E_YIs determined (step S5).
5), all registered data E stored after the registration area_{Science fiction}
Is moved to the destination of one data E to make the registration area empty,
Registration target data E_YIs stored in the registration area and
Step S56), number of data n_{Science fiction}To (n_{Science fiction}+1)
(Step S57), number of updated data n_{Science fiction}And spare
Data group G _{Science fiction}To the spare area A_{Science fiction}To be stored. The registration unit 12 stores the number of updated data n_{Science fiction}
To spare area A_{Science fiction}By comparing with the data storage capacity N of
And spare area A_{Science fiction}New data E_YTo register
It is determined whether or not there is room (step S58).
Number n_{Science fiction}Is smaller than the data storage capacity N and there is a margin.
Finishes the registration process without executing the re-editing process
Is the number of data n_{Science fiction}Is equal to the data storage capacity N,
If it does not exist, the re-editing unit 13 is started, and
Then, a reediting process is executed (step S59). On the other hand, the registration unit 12_{Science fiction}Extract from
Top data E_SfHAnd registration target data E_YAnd compare
As a result, condition E_Y<E_SfHIs satisfied (step S5
4), search target data E_XIs the basic data area A_BfStored in
And the basic data area A_BfTo access
Data number n_{Science fiction}And basic data group G_BfExtract and binary search
The registration target data E_YDetermine the registration area for storing
(Step S510). Condition E_B1H> E_YIf
(Step S511), registration target data E_YIs Basic Day
Group G_B1Start data E_B1HAs the basic data area A
_B1Of the registration target data E_YClimbing
After determining the recording area, the head data area A_HAccess
And start data E_B1HTo be registered E_YUpdate to
(Step S512). Registration target data E_YRegistration area is determined
Then, the registration unit 12 registers all the registrations stored after the registration area.
Data E_BfTo the registration area by moving
After opening, the registration target data E_YIs stored in the registration area
(Step S513). The registration unit 12 stores the registered data E_BfOne
As a result of moving to the destination of the data E, the registered data E_BfBut
Basic data area A_BfTo see if it overflowed (step
Step S514), the number of data n_BfIs less than the data storage capacity N
Overflowing registered data E_BfIf not, the extracted data
Number n_BfTo (n_BfNumber of updated data after updating to +1)
n_BfAnd basic data group G_BfTo the basic data area A_BfCase
To complete the registration process, but the number of data n_BfIs the data case
Overfilled registered data E equal to the storage capacity N_BfWhen comes out
Also has a spare area A_{Science fiction}Access to stored data
Number n_{Science fiction}And preliminary data group G_{Science fiction}And extract the preliminary data
G_{Science fiction}All registered data E stored in_{Science fiction}To one data E
After moving to the destination and leaving the first area, the basic data area
A_BfData E overflowing from_BfIs stored in the first area and
(Step S516), number of data n_{Science fiction}To (n_{Science fiction}+1)
Updated (step S57), the number of updated data n_{Science fiction}And
And preliminary data group G_{Science fiction}To the spare area A_{Science fiction}To be stored. The registration unit 12 checks the updated state as described above.
Number of data n_{Science fiction}To spare area A_{Science fiction}And data storage capacity N
By comparison, the spare area A_{Science fiction}New data E
_YTo determine whether there is room to register
S58), number of data n_{Science fiction}Is less than the data storage capacity N
If exists, register without executing re-editing process
The processing ends, but the number of data n_{Science fiction}Is the data storage capacity N
If they are equal and there is no room, start the reediting unit 13
Then, a reediting process is executed as shown in FIG.
Step S59). On the other hand, the registration unit 12 stores the final data storage area A
_TFinal data E extracted from_TAnd registration target data E_YWhen
(Step S51), the condition E_Y> E
_TIs satisfied (step S52), the registration unit 12
Is the first data area A_HTo access the last head data E
_BfHAnd extract the final data E_TIs the basic data
Data area A_BfAnd spare area A_{Science fiction}(Step S5)
17), registration target data E_YBasic data area for storing
A_BfAnd spare area A_{Science fiction}To determine. Next, the registration unit 12 stores the basic data area A_BfTo
Number of data accessed and stored n_BfAnd basic data
Data group G_BfIs extracted and the number of data n_BfAnd data storage capacity N
Is compared with the basic data area A_BfRegistered vs
Elephant data E_YAnalyze whether there is room to register
(Step S518), the number of data n_BfIs the data storage capacity
If it is confirmed that there is a margin below the amount N,
Registration target data E_YTo the basic data group G_BfRegister at the end of
(S519), and the number of data n_BfTo (n_Bf+1)
After updating (step S520), the number of updated data n
_BfAnd basic data group G_BfTo the basic data area A_BfStored in
To end the registration process, but the number of data n _BfIs data storage
Basic data area A equal to capacity N_BfData to be registered in
E_YIf you confirm that there is no room to register
(Step S518), the next corresponding spare area A_{Science fiction}
And the number of stored data n_{Science fiction}And spare
Data group G_{Science fiction}And the registration target data E_YThe preliminary day
Group G_{Science fiction}(S521) and the data
Number n_{Science fiction}To (n_{Science fiction}+1) (step S5).
7), number of updated data n_{Science fiction}And preliminary data group G_{Science fiction}To
Reserved area A_{Science fiction}To be stored. The registration unit 12 stores the number of updated data n_{Science fiction}
To the spare area A_{Science fiction}By comparing with the storage capacity of
Area A_{Science fiction}New data E_YThere is room to register
It is determined whether or not to perform (step S58), and the number of data n_{Science fiction}
Is re-edited especially if the data storage capacity is less than N and there is room
The registration process ends without executing the process.
Number n_{Science fiction}Is equal to the data storage capacity N and there is no room
If so, the reediting unit 13 is started, and the
An editing process is executed (step S59). As is clear from the above description, the present embodiment
According to the new data E_YThe registration unit 12 for registering
First, the final data storage area A of the file 40_TAnd the beginning
Data area A_HTo access the registration target data E_YClimb
Basic data area A to be recorded_BfAnd spare area A_{Science fiction}Decide
And then the spare area A_{Science fiction}Access to registration data
E_YTo the basic data area A_BfAnd spare area A_{Science fiction}Any of
It is determined whether to register, and the spare area A_{Science fiction}Should be registered with
If specified, the preliminary data group G_{Science fiction}Data to be registered in
TA E_YAre arranged in ascending order, and the spare area A_{Science fiction}Stored in the base
This data area A _BfIf it is determined that it should be stored in
Data area A_BfTo access the basic data group G_BfRegister with
Target data E_YAre arranged in ascending order and then the basic data area A_Bf
In the basic data area A_BfData overflowing
E_BfOccurs again, the spare area A_{Science fiction}Access
And overflowing data E_BfTo the preliminary data group G_{Science fiction}Array in ascending order
After that, the spare area A_{Science fiction}At the end and, if necessary, final
Data storage area A_TTo access the final data E_TUpdate
New or top data area A_HTo access the first
Data E_B1HWill be updated, and the file 40
The number of access times is only 4 to 7 times, and the search time is reduced
Is done. Next, the file re-editing process according to the present invention
Will be described with reference to FIGS. 2, 3, and 6. FIG. 2 and 3
In FIG. 6 and FIG._YTo
Reserved area A _{Science fiction}As a result, the number of data n_{Science fiction}Is the data case
Spare area A _{Science fiction}New day to
TA E_YWhen it is determined that there is no room to register
Then, the reediting unit 13 is started as described above. When the reediting unit 13 is started, the data number n_{Science fiction}
Is equal to the data storage capacity N._{Science fiction}Compatible with
Basic data area A_BfBasic data area A immediately after_BgLess than
All descending basic data areas A_BAccess to each
Basic data group G stored_Bg(However, g = f + 1) or less
All basic data group G of descending_BTo one basic data area A _B
Move to the destination and store it in the basic data area A_BgEmpty
(Step S61 in FIG. 6). Next, the reediting unit 13 sets the spare area A_{Science fiction}Access
Spare data group G accessed and stored_{Science fiction}Extract the new
Tana basic data group G_BgBasic data that became free space as
Area A_BgTo the spare area A_{Science fiction}Empty (step
S62). Next, the reediting section 13 sets the spare area A_{Science fiction}Immediately after
Reserved area A_SgAll subsequent spare areas A_SAccess to
Spare data group G stored respectively_SgAll subsequent
Spare data group G_STo one spare area A_SMove to
Reserved area A_SgIs empty (step S63). Next, the reediting unit 13 sets the start data area A_H
, And all the stored leading data E_BfH
And moved in steps S61 to S62.
Basic data group G_BgStart data E corresponding to the following_BgHOr later
Is updated (step S64). Thus, the basic data
Area A_BfContains the number of data n_BfIs equal to the data storage capacity N
New basic data group G_BfIs stored, but the corresponding
Area A_{Science fiction}Are all free areas (n_{Science fiction}= 0) and new
When registering data E, the spare area A_{Science fiction}Can be registered to
And basic data area A_BgSo far, the reserve area
A_{Science fiction}Number of data stored in n_{Science fiction}Is the data storage capacity N
Spare data group G equal to_{Science fiction}Is the number of data n _BgIs the data case
Basic data group G equal to storage capacity N_BgStored as
But the corresponding spare area A_SgAre all free areas (n_Sg=
0), and when registering new data E thereafter,
Area A_SgCan be registered. As is clear from the above description, the present embodiment
According to the registration unit 12, the data E_YTo the spare area A_{Science fiction}Register with
As a result, the number of data n_{Science fiction}Is equal to the data storage capacity N.
And new data E_YI can't afford to register
In this case, the reediting unit 13_BAnd forecast
Area A_SBasic data group G registered in_BAnd preliminary data
Group G_STo one basic data area A_BAnd one spare area A_S
Each spare area A_SNew day in
TA E_YCan be registered. FIGS. 2 to 6 show only one embodiment of the present invention.
This is only an example. For example, each data E in the data group G, or
Is the first data area A_HEach head data E in_BfHAre in ascending order
It is not limited to what is classified, it is classified in descending order
Many other variations are considered, such as
The effect of the present invention does not change. The structure of the file 200
Is not limited to the file 40 shown in FIG.
Many variations are taken into account, but in each case
The effect remains the same. In addition, the information processing system
The stem 100 is limited to the illustrated information processing system 10.
The file 40 is not specified.
Optical disk drive or floppy disk drive other than the disk drive (DK) 3
Considering many other deformations such as mounting on a pi disc device
However, the effect of the present invention does not change in any case. [0071] As described above, according to the present invention, the information processing system
In the system, the data search range is a pair of data areas
Search time is greatly reduced because it is limited to
And create free space to register new data
Registration time because the number of data to be moved is limited.
Is greatly reduced, and the information processing system
Data retrieval and registration efficiency is greatly improved.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 illustrates the principle of the present invention. FIG. 2 illustrates an information processing system according to an embodiment of the present invention. FIG. 3 illustrates an example of a file structure in FIG. FIG. 4 is a diagram showing an example of a data search process in FIG. 2 FIG. 5 is a diagram showing an example of a new data registration process in FIG. 2 FIG. 6 is a diagram showing an example of a file re-editing process in FIG. FIG. 8 shows an example of a conventional information processing system. FIG. 8 shows an example of a sequential search file structure in FIG. 7. FIG. 9 shows an example of a sequential search process in FIG. FIG. 11 shows an example of a binary search file structure. FIG. 11 shows an example of a binary search process in FIG. 7. FIG. 12 shows an example of a hash search file structure in FIG. Diagram showing an example of hash search processing [ DESCRIPTION OF SYMBOLS 1 Processing unit (CPU) 2 Storage unit (MM) 3 Magnetic disk unit (DK) 4, 40, 200 File 5 Magnetic disk control unit (DKC) 6 Input / output unit (IO) 7 Input / output control unit ( IOC) 10, 100 Information processing system 11 Search unit 12 Registration unit 13 Reediting unit 201 Top data area 210 Data area 220 Spare area 300 Search unit 400 Registration unit 500 Reediting unit

Continuation of the front page (56) References JP-A-62-276639 (JP, A) JP-A-3-263138 (JP, A) Masami Yamatani, Introduction to File Organization, Ohmsha, Japan, July 25, 1980 Sun, 1st edition, p. 65-p. 83 A. V. Eiho J. E. FIG. Hop Croft J. D. Ullman, Data Structures and Algorithms, Japan, Baifukan Co., Ltd., March 10, 1987, first edition, p. 323-p. 336 (58) Fields investigated (Int. Cl. ⁷ , DB name) G06F 12/00 G06F 17/30

Claims (1)

(57) [Claim 1] In an information processing system for retrieving data stored in a file provided in an external storage device, the file can be stored and extracted with a single access.
A plurality of data areas having a data storage capacity of a large number of data are provided, and one spare area having the same data storage capacity as the data area is provided for each of the data areas, and in each of the data areas, A plurality of data classified in the order of magnitude are stored, and when the storage area of each data area is insufficient, a spare area corresponding to the data area is classified in a predetermined magnitude through the data area and the spare area. Only the top data located at the top of the data stored in each data area in the descending order of magnitude is stored in common with each of the data areas and the spare area, and the top data sorted in descending order of magnitude. An area is provided, and when a search request for required data is input to the information processing system, the head data area is searched, and search target data is stored. Search means for detecting a pair of a data area and a spare area that have been searched, and searching for the search target data from the detected data area or the spare area; and a request for registering required data is input to the information processing system. Registration means for searching for the head data area, detecting a pair of a data area and a spare area in which data to be registered is to be stored, and storing the pair in the detected data area or the spare area in the descending order; When there is no room to further register additional data in the pair of the area and the spare area, the data stored in the spare area of the pair of the data area and the spare area which has no room is replaced with a new data area and a spare area. After moving to the data area of the area pair, all of the data area and the spare area including the pair of the new data area and the spare area are moved. Data re-editing means for re-editing the data stored in the data area in the descending order of the size and adding head data of a new data area to the head data area and classifying the data in the order of the magnitude. .