JPS62248031A - Data managing method - Google Patents

Abstract

PURPOSE:To shorten the retrieval time by executing a rearrangement of a data, when a retrieving request is outputted to a data belonging to the rear part of a list structure of a data group, and receiving the retrieving request of the next time and thereafter, in a state that the data is in front of the list. CONSTITUTION:A retrieving instruction of a data, and a value used for identifying the data are supplied to a retrieving instruction register 1 and a retrieving data register 2, respectively, and a value of a link operating counter 17 is initialized to zero. Subsequently, a store address of the data which is used as an object of comparison is obtained in an address register 12, and this data is stored in an identifying part register 13, a link part register 14, and a data part register 15, respectively. Next, a value of the counter 17 is increased by '1', values of the register 2 and the register 13 are compared, and whether the data is that which is to be derived or not is decided. If it is not the data to be derived, a pointer is renewed, a store address of the data concerned is stored in a history register 11, a value of the register 12 is updated to a value of the register 14, and the processing is returned to an input of the data.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a data management method for efficiently performing data searches.

[Prior Art] Data may be searched for a data group having a list structure using sequential search from the head of the list. For example, in a disk cache system, the storage address in the data memory of a data block, which is the unit of data transfer between the disk and the disk cache writing system and is the minimum unit for data management in the disk cache system, is stored in a list structure. An example of this is when managing data as a group of data.

FIG. 12 is an explanatory diagram for explaining a disk cache as an example of managing a data group having a list structure.

In other words, the size of the data processor is 4096 (=163) words, and the capacity of the data memory is 2.
This is an example in a disk cache O system of 0480 (=5×163) words.

(A) is an explanatory diagram showing the disk and disk cache states. For example, the disk cache (+00
) data memory (101) 0000(h)
Disk (+02) from address to address 0OFF(h)
Data block A starting from address A above is stored. Note that (h) indicates hexadecimal display.

CB) is an explanatory diagram showing the state of the disk cache 79, and shows a data group of control information.

In the figure, the control information consists of the storage start address of the data block on the disk, the storage start address of the corresponding data block on the data memory for giving a list structure to the data group of the control information, and the control information memory (10
It is stored in an area called 3). Further, the address of the first data in the list created by the control information data group is held in a register called a list head register (104).

For example, the beginning of the control information data group is stored at address 1000 (h) of the control information memory (103), and data block A starting from address A on the disk (102) is stored in the control information memory (103). Cash (
100) is stored from address oooo(h) of the data memory (101).

In the disk cache example shown in FIG. 12 above, if the data contained in data block C starting from address C on the disk is required, the disk cache system first stores the data at the beginning of the list pointed to by the list head register. The control information of the element is checked, but since it is the control information regarding data block A, it is not the desired information.

Therefore, following the link of the control information, 100
6 (h) The control information at the address is checked, but since it is the control information regarding data block B, it is not the desired control information. So, I followed the link further and 1004
(h) Check the control information for the address, find out from the value of address C that the control information is related to data block C, and confirm that the desired data block is stored from address 0200 (h) in the data memory of the disk cache. know that As in the example in Figure 12 above, when searching for data using sequential search from the top of the list for a data group with a list structure, the desired data points to the first element of the list. It can be reached by a small number of link operations, counting from the button and wrist head. That is, the closer the wrist head is, the faster data can be obtained. Therefore, in order to improve the efficiency of data retrieval, it is necessary that the data to be retrieved be as close as possible to the list head.

A conventional data management method that can meet the above requirements will be described with reference to FIG. 8 and subsequent figures.

FIG. 8 is a configuration diagram showing an example of a data management device for implementing the above data management method.

In the figure, a data management device 7i (80) stores a search command for data managed by a data management type fi (80) in a search command register (81) and data to be searched in a search data register (82). This is a device that stores values for identifying the data as human information and outputs the required data to the output register (82).

(20) is a memory, and this memory (20) stores a data group to be managed, and the data included in the data group is what the data requires, as shown in the figure. an identification section (21), which is a field used for comparison to identify whether the
, and a data section (23) which is a field storing data to be output.

Reference numeral (10) is a list head register, and this list head register (10) holds the storage address of the data that is the first element of the list created by the data group.

(11) is a history register, and this history register (11) stores data used when reconstructing the list structure during search processing.

Also, (12) is an address register, and (13) is
The identification part register (14) is the link part register (1
5) are data part registers, and these registers are used to store the address, identification part, link part, and data part of the data being compared at that time, respectively, during the search process. It is.

Next, the operation of the data management device will be explained with reference to the flowchart shown in FIG.

When a data search command is given to the search command register (81) in FIG.
) is given a value used to identify the data and a search begins.In step (30), the storage address of the data to be compared is obtained in the address register (12).

Next, step (31) as shown in detail in FIG.
, the identification part, link part, and data part of the data at the storage address specified in the address register (12) are stored in the identification part register (13) and link part register (14), respectively.
) and stored in the data section register (15) (step 3
11,312,313).

Next, the value of the search data register (82) and the value of the identification unit register (13) are compared to determine whether or not the data currently being compared is what is being sought. If the data is not the desired data, the process proceeds to step (32) via branch (36).

In step (32), as shown in detail in FIG. 4, in step (321), the history register (11
), store the storage address of the data and stay.

Step (322) updates the value of the address register (12) to the value of the link register (14), and step (31) updates the value of the address register (12) to the value of the link register (14).
), but since the storage address of the data, which is the next element of the list structure, is specified in the link part of the data included in the data group, the address register (12) is
Storage address of data to be compared, history register (11
), the storage address of the data compared immediately before is stored, and the processing from step (31) is repeated.

On the other hand, if the relevant data is the first data, branch (35)
Next, the value of the address register (12) and the value of the list head register (10) are compared to determine whether the data concerned is the first data of the list structure. If the data is the first data, branch (38
), performs the processing of step (33), and then proceeds to step (34), outputs the value of the data part register (15) to the output register (3), and ends the search. However, in step (33), as shown in detail in FIG. 5, in step (331), the link part ( 22) to the value of the link part register (14), and in step (332), the link part (22) of the data whose address is the value of the address register (12) is changed to the list head in the memory (20). The value of the register (lO) is changed, and in step (333),
The value of the list head register (10) is transferred to the address register (
12).

That is, the list structure of the data group in the memory (20) is reconfigured, and the data before and after the desired data are continued after the data.

[Problems to be Solved by the Invention] Since the data management device for implementing the conventional data management method is configured as described above, it is difficult to retrieve data in response to a search request according to a certain pattern. There was a problem in that efficiency was poor due to unnecessary sorting. This problem will be explained in more detail with reference to the drawings.

FIG. 10 is a definition diagram for simply expressing a list structure having a data group.

FIG. 8A is a detailed diagram of the list head register (lO) and memory (20) of the data management device (80) in FIG. 8, and shows the values of data stored in each.
represents an example.

In FIG. CB), head (50) represents the list head register (10). Also, arrow (60)
indicates that the value of the list head register (10) is the storage address of data whose identification part is 8;
In case A), the value of the link register (10) is 1.
000 (h).

The node (51) represents data whose identification part is A, and in FIG.
(h) Represents data for location i. Notes below (
The same applies to 52), (53), (54), and (55).

Further, the arrow (61) indicates that the value of the link part of the data whose identification part is A is the storage address of the data whose identification part is B. In FIG. 20
), the link part of the data at address 1000(h) is 1006
(h). Arrows (62), (63
) and (64) are also similar.

Tail (56) represents the terminal symbol of the list created by the data group, and in FIG. 10 (A), memory (20)
This is the data at address 1008(h).

The arrow (65) indicates that the link part of the data whose identification part is E represents the storage address of the terminal symbol data, and in FIG.
This indicates that the value of the link part in the address data is (tooth).

Hereinafter, the notation shown in FIG. 10(B) will be used to express the list structure constituted by the data group.

Further, unless otherwise specified, data whose identification part is A will be referred to as A in width.

Here, a data group consisting of A, B, C, ...-9Z is stored in a data management device using a conventional data management method.
Assume that the list structure is stored as shown in FIG. 11(A). When the data management device is installed in this state, there are search requests that repeat the pattern of "search for A, then search for B, then search for C", that is, A, B, C, A, B, C,
The flow of processing when search requests A, B, C, . . . are changed will be described below.

First, given a search request for eight, first,
The first node in the list is examined, but as shown in Figure 11 (a), the data of this node is C, not A. Therefore, follow the link of C according to branch (36) in Figure 9. , then examine the second node. However, the data of this node is B, not A. Therefore, the link of B is followed to check the third node. The data of this node is A, which is the desired data, and this node is not at the beginning of the list. Therefore, the data is rearranged according to branch (38) in Figure 9, and A is moved to the beginning of the list. .

In this way, in response to the search request for A, the data management device finds A in the third node of the list and arranges A at the top of the list, as indicated by diagonal lines in FIG. change. As a result, the list structure of the data group is as shown in Figure 11.
will be reorganized as follows.

Next, when a search request is made for B, as above,
As shown in Figure 11 (1), B is found in the third node of the list, and it is rearranged to the top of the list, and the result is as shown in Figure 11 (1). Furthermore, a search request for C is made. , as shown in FIG. 11(1), C is found at the third node and rearranged to the top, resulting in the result shown in FIG. 11(e).

In the state shown in FIG. 11(E), a search request for A is issued again, but the state shown in FIG. 11(O) is different from that shown in FIG.
Since the situation is the same as in b), A, B, and C after the second mouth are
The search process is a repetition of the first process.

The problem here is that any data A, B, or C is discovered at the third node in the list.

In Figure 11 (a), B is the second node of the list,
Further, C exists at the first node. Therefore, if the data had not been rearranged, the data management device would have found B as the second node and C as the first node.

In conventional data management methods, data is rearranged, which leads to a decrease in search efficiency, and this inefficiency is reproduced every time the search pattern of "A, B, C" is repeated. There was a problem that

This invention was made to solve the above-mentioned problems, and aims to provide a data management method that eliminates the inefficiency of data retrieval caused by rearranging data.

[Means for Solving the Problems] In the data management method according to the present invention, when searching for certain data, the number of times links in a list structure created by a data group are followed, which is the number of link operations during the search process, is determined. When data is found, the number of link operations required during the search process is compared with a certain predetermined value called the boundary value, and if the number of link operations is less than the boundary value, , the data is not rearranged.

In other words, for a data group with a list structure, depending on the boundary value, a set of data that can be reached from the beginning of the list by a number of link operations equal to or less than the boundary value is called the front part, and a set of other data is called the rear part. When searching for data, the data in the front part is not rearranged, but the data in the rear part is rearranged.

[Function] In the data management method of the present invention, when a search request is issued for data belonging to the rear part of the list structure of a data group, the data is rearranged so that subsequent search requests can be made The search time is shortened by receiving the file at the front of the list.

Further, when a search request is issued for data that bends toward the front, the data is not rearranged, and unnecessary rearrangement that causes a decrease in search efficiency is omitted.

[¥Example] An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an example of a data management device for implementing the data management method of the present invention.

In this embodiment, a link operation counter (17) is used to count the number of link operations during data search processing.
and a boundary setting register (1
6) is provided in the data management device (100).

Note that other components in FIG. 1 are the same as those in FIG. 8.

Next, the operation will be explained.

FIG. 2 is a flowchart showing details of processing performed using the data management device whose configuration is shown in FIG.

Therefore, when a data search command is given to the search command register (1) and a value used for data identification is given to the search data register (2) and the search begins, step (4)
0), first, the link operation counter (17) is initialized to zero.

Next, in step (30), the address register (
12) Obtain the storage address of the data to be compared.

Next, as shown in FIG. 3 in detail, Stella 7' (
31), the identification part, link part, and data part of the data at the storage address specified in the address register (12) are
They are stored in the identification section register (13), link section register (14), and data section register (15), respectively (steps 311, 312, and 313).

Next, in step (41), the value of the link operation counter (17) is increased by one.

Next, the value of the search data register (2) and the value of the identification unit register (13) are compared to determine whether or not the data currently being compared is what is being sought. If the data is not the desired data, the process goes to branch (46) and the process of step (32) is performed.Here, as shown in detail in FIG. Step 321) to store the storage address of the data in the history register (11), step 322) to update the value of the address register (12) to the value of the link register (14), and step 322) of step (31). Returning to the process, since the storage address of the data, which is the next element of the list structure, is specified in the link part of the data included in the data group, the storage address of the data to be compared and the history are stored in the address register (12). The storage address of the data compared immediately before is stored in the register (11), and the processing from step (31) is repeated.

On the other hand, if the data is the desired data, step (
45), the value of the link operation counter (17) and
The values of the boundary setting register (16) are compared and it is determined whether the data belongs to the front part or the rear part of the list structure.

Then, if the data is in the front part, branch (
47), and in step (34) the data section register is
The value of (15) is output to the output register (3) and the search is completed. However, in step (33), as shown in detail in FIG. The data is connected with a link part, the data is made the first element of the list, and the data that was at the beginning is continued after the data (step 33).
1,332,333).

In summary, this embodiment differs from a data management device using a conventional data management method by counting the number of link operations during data search processing and by comparing the number of link operations with a certain standard. The process is divided into cases where data is sorted and cases where data is not sorted.

Next, how the data retrieval efficiency in the data management method of the present invention described above is improved by the present invention will be explained with reference to the drawings.

FIG. 6 is a definition diagram of an abbreviation for simply expressing the structure of a data group managed by the data management method of the present invention.

Figure 1 (A) shows the list head register (10), boundary setting register (16), and memory (20) extracted from the data management device in Figure 1, and shows an example of the data stored in each. It represents.

In FIG. CB), a vertical line (70) represents a boundary that divides the data group into a front part and a rear part.

In the example of FIG. 6(A), the value of the boundary setting register (16) is 0003(h), so the three data A, B, and C from the beginning of the list are the front data, and the remaining data and E are Since this is rear data, the boundary (70) is set between C and 0.

Note that the other components in FIG. 6 are the same as those shown in FIG. 10.

Now, in a data management device using the data management method of the present invention, a data group consisting of A, B, C, ..., Z is stored in a list structure as shown in FIG. Suppose that

In the data management device in the above state, repeating "A, B, C", that is, A, B, C, A, B, C, A, B, C
, ---..., and the processing is as follows.

First, when a search request for A is issued, A is found in the third node of the list, as indicated by diagonal lines in Figure 7 (a), but since A is data in the front part, the conventional management method is used. Unlike , data is not sorted.

In response to the next search request for B, the data management device (100) retrieves B from the 27th search request of list I, as shown in FIG.
- Discover it quickly. In this case as well, data is not rearranged.

In response to the next search request to C, the data management device 1Ii
(100) is now C as shown in Figure 7 (1).
is found in the first node of the list.

In this case as well, the data is not rearranged as in the above case.

When searching for "A, B, C" for the second time, the above process is repeated. In this example, the data has never been sorted, and in the conventional data management method, B, which would have been discovered in the third node due to unnecessary sorting, C, respectively 27th
- Can be found at the front mate.

[Effects of the Invention] As described above, according to the present invention, wasteful data rearranging operations can be omitted when searching for data, and data that was generated due to data rearranging in conventional data management methods can be eliminated. It is possible to suppress search inefficiency, and
The data management method of the present invention has effects such as being extremely effective in cases where a certain pattern of data search requests is repeated as in the above embodiment.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a data management device for carrying out the data management method of the present invention, FIG. A flowchart showing detailed contents of the data import process in the flowcharts of FIGS. 2 and 9 above, and FIG. 4 are similar to those shown in FIGS.
FIG. 5 is a flowchart showing the detailed content of the pointer reordering process in the flowchart of FIG. 2 and FIG. 9, and FIG.
, (B) is a definition diagram showing the abbreviation of the data structure in the data management method of the present invention, and FIGS. 7(a) to 7(e))
8 is an explanatory diagram showing the data search procedure in the data management method of the present invention, FIG. 8 is a block diagram of a data management device for implementing the conventional data management method, and FIG. 9 is the same as that shown in FIG. 8 above. 10 (A) and 10 (B) are flowcharts showing processing procedures for data retrieval in a data management device. FIGS. , 11-T is an explanatory diagram showing the data search procedure in a conventional data management method, and FIGS. 12(A) and 12(B) are disk caches as an example of managing a data group having a list structure. (1)...Search command register, (2)...Search data register, (3)...Output register, (10)...List head register, (11)...・・・
History register, (12) Address register, (13) Identification unit register. (14)... Link section register, (15)... Data section register, (16)... Boundary setting register, (17)... Link operation section counter, (20)...
Memory, (100)...data management device. In each figure, the same reference numerals indicate the same or corresponding parts.

Claims (1)

[Claims] In order to give a list structure to a set of data, a new field called a link is added to each data, and each data link stores the address of the data that should be located next to it in the list structure. A predetermined specific value called a terminal symbol is stored in the link of the data that should be the last one.
A boundary value having a certain value is set, and the set of data that can be reached by following links from the beginning of the list a number of times less than or equal to the boundary value is defined as the forward part, and the other set of data is defined as the backward part, When searching for data by dividing the list into a front part and a back part, the data is sequentially searched from the beginning of the list, and if the desired data is found in the front part of the list, it simply outputs the requested data. , a data management method characterized by not only outputting the requested data but also reorganizing the list structure and moving the found data to the front part of the list when the data is found in the rear part of the list.