JPH02227735A - Name table retrieving device - Google Patents

Abstract

PURPOSE:To shorten the searching time of labels despite the increase of labels by storing a chain of labels stored in a name table and having the same hash value into a hash table of the name table in a Mr. B-tree structure. CONSTITUTION:A label register means 12 takes labels one by one out of a definition work file 11 storing the labels at random and stores a chain of labels stored in a name table 13 and having the same hash value into a hash table of the name table 13 in a Mr. B-tree structure. At the same time, a name table retrieving means 15 searches the labels chained in the Mr. B-tree structure out of the hash table based on the labels stored in a reference work file 14. Thus it is possible to define the time required for retrieving as the value propor tional to log(m/n), where (n) and (m) mean the hash width and the number of component elements, respectively. As a result, the label retrieving time is shortened despite the increase of labels.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention is used for managing a name table of a language processing system in information processing. The present invention relates to searching name tables in language processing systems.

[Winter required]

The present invention is a name table search device that uses the chain hash method to register names on a name title and searches for them, in which a chain of names in a name table having the same hash value is stored in a three-name tree structure. However, by searching for name marks that are chained in a three-family tree structure, the time required for name name search can be reduced even if the number of stored name names increases.

[Conventional technology]

Conventionally, this type of name table search has used a chained hash method that uses a Hatch function that is uniquely determined from a name name to chain names with the same Hatch function from a hash table with a finite size. .

[Problem that the invention seeks to solve]

In the conventional chain hash method described above, even when an ideal hash function is assumed, where the width of the hash value is n and the number of constituent elements is m, the order of time required for search is proportional to m/n. Therefore, as the number of landmarks increases, the search efficiency decreases significantly.

The present invention aims to eliminate such drawbacks, and aims to provide an apparatus that can reduce the time required for searching even when the number of landmarks increases.

[Means for solving problems]

The present invention provides name mark registration means for extracting name marks from a definition work file in which name marks are stored in random order and registering the name marks on a hash table included in a name table using a chain hash method, and name marks to be referenced. and a name table search device for searching for a name registered in the name table according to the name of a reference work file in which the name is stored, the name table search device comprising:
comprising means for storing a chain of names in the name table having the same hash value in an 8-tree structure in the hash table; The invention is characterized in that it includes means for searching for famous landmarks.

Here, the 8-tree structure is an organization that dynamically maintains indexes as data changes and guarantees appropriate data retrieval efficiency and storage space utilization at any time, and is preferred for database applications. (Virtual S
It is also known as the basic concept of storage access method. (Refer to Information Processing Handbook: Edited by Information Processing Society of Japan) [Operation] The name registration means extracts the name from the definition work file and creates a chain of names in the name table with the same hash value in the hash table on the name table. Stored in an 8-person tree structure. A name table search means searches the hash table for name names chained in an eight-name tree structure according to the name names stored in the reference work file.

Thereby, even if the number of stored names increases, the time required to search for the name can be shortened.

〔Example〕

Next, embodiments of the present invention will be described based on the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, and FIG. 2 is a diagram showing an example of a chain of name tables 13 according to the embodiment of the present invention.

The embodiment of the present invention extracts name marks from a definition work file 11 in which name marks are stored in random order, and stores them in a hash table 2 included in a name table 13 using a chain hash method.
1, and a name table search means 15 for searching for a name registered in the name table 13 according to the name in a reference work file 14 in which names to be referred to are stored. As a feature of the present invention, the name registration means 12 includes means for storing a chain of names in the name table 13 having the same hash value in the hash table 21 in an 8-family tree structure, and the name table search means 1
5 includes a means for searching the hash table 21 for names that are chained in an eight-family tree structure.

Next, the operation of the embodiment of the present invention configured as described above will be explained.

Name tags are stored in the definition work file 11 in random order, and the name table registration means 12 takes out the name tags one by one from the definition work file 11 and registers them in the name table 13.

FIG. 2 is a diagram showing the name table 13 according to the embodiment of the present invention. The hash table 21 is based on the badge μ function (
(takes a value from 0 to n) to obtain a hash value n1, and store the name name in an array of origin pointers for chaining from the entry corresponding to nl. It takes a NULL value when no constituent elements are chained.

Each component of the name table 13 is a hash table 2.
1 to 22 components (in the case of = 1), the pointer to the left of each name points to a component with a name smaller than or equal to that name. , the pointer to the right points to a component with a name larger than that name.

Also, if there is nothing to point to, the value is NULL.

In the name table 13 of this format, the name name ABC<BCD<
BCE<BCG<BDE<CDE<DEF<EAB<E
Chain examples 23 to 30 are examples in which FG<FGH<FGJ are chained (assuming that they use a hash function that has the same hash value).

The name name registration means 12 in the name table 13 registers the name name in the name table 13 shown in FIG. 2 by the process shown in FIG. In FIG. 3, the process of determining the position n1 of the hash table 21 from the name tag read from the definition work file 11 and registering it as an element of the B mature tree is repeated for all the name tags.

That is, one name tag is read from the definition work file 11 (step 302), it is determined whether it is finished or not (step 303), if it is not finished, the hash value n1 is calculated from the name name (step 304), and the hash table 21 n of
Step 3: Register the name tag in l as an element of the node of B mature tree.
05), repeat for all name tags, step 30
If the determination in step 3 is completed, the process ends (step 306). The name table search means 15 uses the B mature tree to search the name table 13 for the names in the reference work file 14 as shown in FIG. FIG. 4 shows a procedure for extracting a landmark from the reference work file 14, determining its hash value, and searching for a B mature tree from the chain start point on the hash table 21.

That is, step 402) reads a reference name from the reference work file 14, it is determined whether or not it is finished (step 403), and if it is not finished, a hash value n2 is determined from the name (step 4θ4). Next, it is determined whether the hash value n2 of the hash table is NULL or not (step 405), and if it is NULL, the process is returned as an undefined error (step 410); if it is not NULL, the hash value n2 of the hash table is Search for the B mature tree pointed to (step 407), determine whether it is found or not (step 408), if not found, issue an undefined error and return the process step 410), if found, check to see if it has been defined twice. If there is a double definition, return the process as a multiple definition error (Step 4)
12) If there is no double definition, the name name found as a search result is processed (step 411). Repeat this procedure until the reference name reading is completed.
When it is determined in step 03 that the process has ended, the process ends (step 406).

[Effects of the Invention] As explained above, according to the present invention, by forming the chain of the chain chain and the Sosch method into a three-tree structure, the width of the /'% chain is n, where m is the number of constituent elements. The time required for searching can be set to a value proportional to log(m/n), which reduces the time required for searching when the number of names increases compared to searching a name table using the conventional chain hash method. There is an effect that can be done.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention. FIG. 2 is a format of a name table according to an embodiment of the present invention. FIG. 3 is a flowchart showing the processing flow of the name registration means according to an embodiment of the present invention. FIG. 4 is a flowchart showing the processing flow of the name table search means according to the embodiment of the present invention. 11... Definition work file, 12... Name mark registration means, 13... Name table, 14... Reference work file, 15... Name table search means.

Claims (1)

[Claims] 1. Name name registration means for extracting name names from a definition work file in which name names are stored in random order and registering them on a hash table included in a name table using a chain hash method; a name table search device for searching for a name registered in the name table according to a name in a reference work file in which a name to be registered is stored; a means for storing a chain of names in the name table having a hash value in a B-tree structure in the hash table; A name table search device comprising means for searching.