JPH04257062A - Data processor using hash queue - Google Patents

Abstract

PURPOSE:To suppress the increase of queues at a minimum without using any overflow queue concerning the data processor using a hash queue to decide the registering position of data by applying a hash function as an item key value. CONSTITUTION:By providing a hash processing part 13 to prepare the plural hash functions decided the order of application and to successively apply the next hash function when there is no vacancy at the hash queue calculated by the hash function applied before, the queue to register data is decided by combining the hash functions on multiple stages.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data processing apparatus using a hash queue for converting the key of a data item into a physical storage address according to a specific hash function and registering data at that location. It can be used for managing files related to direct access volumes and similar table management.

0002

2. Description of the Related Art FIG. 4 is an explanatory diagram of the prior art. A hash function based on a specific algorithm is applied to a certain item of data to be registered, and depending on the result, one of the hash queues 15 as shown in (a) of FIG. 4 is selected and the data is registered. is being carried out. When searching for data, the same hash function is used to find the storage location of the data.

In conventional hash-based queue table management, the hash queue 15 has normal queues 15a to 15c.
and an overflow queue 15d, for example, as shown in FIG.
As shown in (b), when the hash processing unit 40 applies a hash function to the item and uses the result as a queue number, data is entered into the normal queue (for example, the normal queue 15a) corresponding to that queue number. If it is not turned off, it is handled by registering it in the overflow queue 15d.

0004

[Problem to be Solved by the Invention] In the prior art, all items that cannot fit in the queue are set as overflow items in the overflow queue 15d. Therefore, if the file management on the direct access volume is in a hash structure, the items that cannot be set in the queue are set in the overflow queue 15d. Since the amount of items to be stored is limited by the track size, there is a problem in that many items are set in the overflow queue 15d.

[0005] Also, in order to reduce the number of overflow items,
Although it is possible to increase the number of queues, this often results in a waste of resources, and it may be difficult to increase the number of queues due to constraints on the amount of resources.

SUMMARY OF THE INVENTION The present invention aims to solve the above-mentioned problems and provides a means for minimizing the increase in the number of queues without using an overflow queue.

[0007]

[Means for Solving the Problems] FIG. 1 is a diagram illustrating the principle of the present invention. In FIG. 1, 10 is a data processing device including a CPU and memory, 11 is a data registration unit that registers data, 12 is a data search unit that searches for registered data, 13 is a hash processing unit, and 14 is a magnetic A data storage device using a direct access volume such as a disk device, 15 represents a hash queue.

When data is registered by the data registration section 11 and when data is searched by the data search section 12, the storage location of the data is determined by the hash processing section 13.

[0009] In the present invention, a plurality of hash functions used by the hash processing section 13 are prepared. The application order of these hash functions F1, F2, . . . is determined in advance. The first hash function F1 is applied to the data item, and the storage position of the data is determined based on the resulting queue number Q1. If there is no room to store an item in queue number Q1, hashing is performed using the next hash function F2 instead of using an overflow queue. Then, the item is stored in the resulting queue number Q2. If there is no space in the queue, hashing is further performed using the following hash function.

0010

[Operation] Multiple hash functions are prepared, and when there is no room in the table to store an item in a queue calculated using a certain hash function, the calculation is performed using a different hash function and the item is set in another queue.

[0011] In this way, by applying different types of hash functions in multiple stages, it is possible to avoid using an overflow queue and to minimize the increase in the number of queues. Therefore, it becomes possible to save resources. Additionally, by first setting data in a queue that can be stored, efficiency during searches can be maintained.

0012

Embodiment FIG. 2 shows an example of application of the present invention. In a direct access volume 20 such as a magnetic disk, a file management list 21 containing management information of files stored in the volume is provided in a number of specific tracks. Note that this information is stored as a VTOC (Volume Table
Of Contents).

[0013] By using one track as one hash queue, file management information is hashed. The key value used to calculate the hash function is the file name, and by associating each track on the direct access volume 20 with a hash queue, the track in which the management information for that file is stored can be immediately determined from the file name. do it like this.

FIG. 3 shows a processing flowchart according to one embodiment of the present invention. Hereinafter, the process (a) shown in FIG.
i) Processing at the time of file registration will be explained according to . Note that in this example, the processing is performed in a loop for convenience of explanation, but the processing is slightly faster if multi-stage processing is realized without a loop. (a) Use the file name to be registered as the key value. (b) Initialize loop variable n to 1. (c) Calculate a predetermined n-th hash function according to the key value of the file name.

(d) The result is the track address of the direct access volume 20 shown in FIG. (e) Determine whether there is room to register file management information on that track. If there is space, proceed to process (f); if there is no space, proceed to process (g).
Proceed to. (f) Register the management information of the specified file in the free area of the track, and end the process.

(g) If there is no free space in the area of the track, it is determined whether the next hash function exists. If there is the next hash function, proceed to process (h). If not, proceed to process (i). (h) Add 1 to n, and repeat the process (c) below. (i) If no empty track is found even after applying all the hash functions, registration is not possible and the process ends with an error.

Although the processing at the time of file registration has been described above, the processing at the time of file search is also similar. In the case of a search, hash functions are calculated in the same order as when the file was registered, using the file name as a key value, until the corresponding file is found, and the track in which the file management information is stored is determined.

[0018]

[Effect of the invention] As explained above, according to the present invention,
There is no need to use an overflow queue, making it possible to use hash queues effectively, and reducing the increase in the number of hash queues themselves, resulting in resource savings.

[Brief explanation of the drawing]

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

FIG. 2 is a diagram showing an example of application of the present invention.

FIG. 3 is a processing flowchart according to an embodiment of the present invention.

FIG. 4 is an explanatory diagram of the prior art.

[Explanation of symbols]

10　　　Data processing device 11 Data registration section 12 Data search section 13 Hash processing section 14 Data storage device 15 Hash queue

Claims (1)

[Claims] [Claim 1] A data processing device (10) that determines a hash queue (15) in which data should be registered by applying a predetermined hash function to an item to be hashed.
, multiple hash functions (
F1, F2, ...), and includes a hash processing unit (13) that sequentially applies the next hash function when there is no space in the hash queue (15) obtained by the previously applied hash function. A data processing device using a hash queue, characterized in that functions are combined in multiple stages.