JPH04340165A - Information retrieval processing system - Google Patents

Abstract

PURPOSE:To speed up data base retrieval processing using multi-keyword and to rapidly execute partial coincidence retrieval to be executed by inputting an optional character string. CONSTITUTION:A retrieval file is previously formed by constituting a character set of each character in a keyword string to be retrieved and plural (r) characters following to the character concerned, forming character set position information constituted of the record identification(ID) code of a keyword including the character set, a character set position order code indicating the character position in the keyword and a keyword attribute code indicating the logical section of the keyword and grouping the positional information in each character set sort. At the time of inputting a retrieval request, each retrieval input is broken down to a character set consisting of (r) characters to constitute a retrieval input character set string, the character position information of he analized character set string is extracted from the retrieval file and collated and a keyword string record having a common record ID, code, equal character order and the same keyword attribute code to/as the retrieval input is extracted from the retrieval file.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information retrieval processing method for performing information retrieval. The present invention is particularly suitable for performing partial match searches using multi-keywords, and can significantly reduce the number of character string matches between input multi-search input strings and registered keywords. The present invention relates to an information search processing method that can perform each search process for a match (middle match, prefix match, suffix match) and complete match at high speed. The present invention is suitable for an information search processing method that performs multi-keyword partial match search processing in a database system.

0002

[Prior Art] Conventionally, strings of keywords possessed by records to be searched are compared with search input strings from the beginning to the end, and keywords that match the input string specified by the searcher and the search conditions are searched. You can use the sequential search method to search for record numbers, or create a string that can be searched from keywords and store the search file in an index format. An index method that searches using an index structure is a common partial match search technique using multiple keywords.

0003

[Problem to be Solved by the Invention] However, in the sequential search method, the search input character string is checked from the beginning to the end of the keywords of the record to be searched, so when there are a large number of keywords, the search requires many It takes time. For this reason, dedicated processors and LSIs that perform high-speed character string matching have been proposed for large-scale database searches, but these methods are limited in hardware and require a computer to perform the search processing and a dedicated processor. and LSI
It takes a long time to transfer character strings between the two systems, and the challenge is to achieve a high speed that satisfies the system.

[0004]Also, although the index method can speed up partial match searches, it has the disadvantage that the search file becomes huge. For this reason, exact, prefix, and suffix searches are used, but intermediate matches are often not supported. This is to do an intermediate match,
In addition to search indexes for exact matches, prefix matches, and postfix matches, a large number of indexes for intermediate matches are required.
The main reasons are that the storage capacity of the search file becomes huge, the search time increases accordingly, and the maintenance of the search file is not easy. Also, depending on the system, search for all prefixes and suffixes of keywords may not be supported due to limitations on the size of the search file. However, searchers often memorize characteristic characters or character strings in keywords, and therefore there is a need for support for partial match searches including intermediate matches to enable smooth searches.

[0005] The inventor of the present invention focused on the fact that the Japanese language has the characteristic that the same characters or the same character strings rarely appear in words that can be used as keywords. Create a character set with that character and the next r characters in total, store these for each character set type and create a search file, and when searching, start with the constituent characters of any character string entered for search. It has been found that searching can be speeded up by decomposing a character into a character set of r characters, extracting the same character set as the decomposed character set from a search file, and checking the continuity of the character set.

[0006] From the above-mentioned viewpoint, the present invention can realize high-speed partial match searches using multi-keywords targeting a huge database system, and is not limited to specific hardware, and can perform search processing. A highly versatile information retrieval processing method that eliminates the need to transfer character strings to and from a dedicated processor or LSI because it is performed in main memory, and allows arbitrary character string searches by focusing on character sets and character set positions. The purpose is to provide

[0007]

[Means for Solving the Problems] The first feature of the present invention is that
record identification code assigning means for assigning a code in ascending order to each record to be searched; keyword attribute code assigning means for assigning an attribute code indicating a logical classification of the keyword to each keyword possessed by this record; Take out one character and add that character and the next sum r
Character set position order code adding means for creating a character set with characters and adding a character set position order code indicating the first character position of the character set in a keyword; the record identifying means; a keyword attribute code; and a character set position order code. The present invention is characterized by comprising means for creating character set position information consisting of the following characters, storing this character set position information in an area for each character set type, and creating a search file.

[0008] Character set position information is about a keyword string created by arranging each keyword of a record in a keyword attribute area corresponding to a keyword attribute code, and all character sets of each keyword are arranged in a keyword attribute area corresponding to a record identification code and a keyword attribute code. It is created by converting the attribute code and character set position order code into a code consisting of integers. Record identification code x n + (Pa -1) + character position order code n: number of keyword string characters Pa: keyword attribute code a. It is preferable that the number is given as the first character position in the keyword string in the keyword attribute area.

A second feature of the present invention is that the search file created according to the first feature is provided, and the characters constituting the search input string are broken down into a character set of r characters from the first character. A means of creating a set string and extracting character set position information of the same character set as this character set from the above search file, and a common record identification code and keyword attribute code between the character set position information of each character set extracted. a means for extracting a combination of character set position information in which the difference in character set position order codes is equal to the first character position difference of the corresponding character set of a search input character string and whose keyword attribute code is equal to the search input; The present invention is characterized by comprising means for outputting a record identification code corresponding to a search input character string as a search result based on the combination of character set position information obtained.

[0010] When a search input string is decomposed into a character set of r characters from the first character, the last character set may be less than or equal to (r-1), and it may not be possible to create a character set of r characters. . At this time, it is preferable to take out the missing number of characters from the rear of the character set immediately before the last character set and connect them to the front of the last character set to create a character set of r characters.

[0011] Furthermore, it is preferable to extract combinations of character set position information that can form the same character set string as the search input character set string in order of character sets that appear less frequently in all keywords of the search input.

[0012] Furthermore, in order to extract combinations of character set position information that can constitute the same character set string as the search input character set string, character set position order codes of character sets that appear less frequently in all keywords of the search input character set string are extracted. i, character set positional order code of character set with high frequency of appearance is j, (character set positional information of character set of character set positional order code i) - (character set of character set of character set positional order code j) It is preferable to use the formula: position information)=ij.

[0013] If the keyword is a Roman character string that includes symbols, it is preferable to use a character set of at least three characters and symbols, and to use a search file for a character set type group containing only European characters that include symbols.

[0014] Furthermore, when the keyword includes a kanji character, a search file can be used that provides character position information in units of one character for kanji characters and character set position information in units of at least two characters for kana characters.

[0015]

[Operation] The frequency that the same character string appears in a document is low. For example, the Kojien dictionary has about 9 million characters in its explanatory text for headwords, but when you look at the frequency of kana characters, the average frequency of occurrence is about 53.
As high as 200 times. However, when we examine the appearance frequency of two-character kana character strings, the average appearance frequency is as low as 472 times. Therefore, if the character set is kana 2 characters, if the search input is n characters, the matching targets to be extracted will be (n/2) on average.
)×472 character set position information. Since kanji has more character types than kana characters, the frequency of appearance of a string of two kanji characters is lower than that of kana characters, and the number of matching targets to be extracted is also smaller than that of kana characters. Generally, a search input is a few dozen characters or less, so even if the string includes characters that appear frequently, the number of times it will be matched will be extremely small compared to a method that sequentially matches all characters.

For example, although there are many terms that use the two character strings ``communications'', the character string ``communications...'' has four characters with the same character, such as ``communication line'' and ``communication device.'' It occurs very infrequently. As a result,"
If you match the character strings ``line'' and ``device'' that follow ``communication'', your search will be narrowed down rapidly. In this way, as the character sets that make up the search input character set string are compared with the search file, keywords that differ from the search input character set string are found among the search target candidate keywords obtained so far. The search target is narrowed down for each constituent character set to be deleted. In particular, if the matching is performed in order of character sets that appear less frequently among all the keywords in the search input, the search will be further narrowed down and the number of matching matches can be reduced.

[0017] Therefore, corresponding to each character set constituting the keyword of the record to be searched, a record identification code that can identify the record, a keyword attribute code that can identify the attribute of the keyword, and a character set position in the keyword are determined. Create character set position information from a character set position order code that can identify the By performing verification and matching, it is possible to significantly reduce the number of times of verification and matching processing in character string searches.

This search file is created as follows.

First, record identification codes are assigned to each record to be searched in ascending order according to the registration order. Next, the logical type of the keyword that each record has is set as an attribute, and a keyword attribute code indicating the attribute is assigned. For example, in the case of a book search system, a keyword attribute code indicating attributes such as book name, author name, publisher name, year of publication, and abstract is assigned. Furthermore, extract one character from the first character of the keyword, create a character set with that character and the next r characters in total, and assign a character set position order code to each character set that indicates the position of the first character of the character set in the keyword. Give. Character set position information for each character set of the keyword is generated from these record identification codes, keyword attribute codes, and character set position order codes. At this time, so that keyword attributes can be represented by character positions, the first character position of the keyword, which is set in advance for each keyword attribute code, is added to the character set position information as a constant. This character set position information is grouped by character set type, and each group is assembled to create a search file. Therefore, this search file has a file structure in which character set position information is stored for each character set type.

In the search process, one or more pairs of search input character strings and search input character string attributes are input. For each search input string, the search input string is decomposed into a character set of r characters from the first character, the character set position information of the same character set as the decomposed character set is extracted from the search file, and the record identification code is common. Then, a combination of character position set position information in which the difference in character set position order code is equal to the first character position difference of the character set of the corresponding search input character string and whose keyword attribute code is equal to the search input is collated and extracted. A record identification code corresponding to the search input character string is extracted as a search result from the combination of character set position information corresponding to each retrieved search input.

[0021] When a search input string is decomposed into a character set of r characters from the first character, the last character set may be less than or equal to (r-1), and it may not be possible to create a character set of r characters. . At this time, characters corresponding to the number of missing characters are extracted from the rear of the character set immediately before the last character set, and are concatenated to the front of the last character set to create a character set of r characters.

[0022] This matching process checks whether the character set strings of the search input and the search file match in continuity and the attributes match, and it is determined whether the records have common identification codes based on the character set position information in the search file. This is done by extracting keywords for which the difference in character set position order code is equal to the difference in the first character position of the character set of the corresponding search input character string and the keyword attribute code is the same.

[0023] This eliminates the need to match all the keywords in the search file, and it is only necessary to match the character set position information of the same character set as the search input in the search file, so the number of matches is reduced to sequential matching. can be significantly reduced compared to . Also, in Japanese words,
Since the frequency of occurrence of the same character string is low, the search target is narrowed down each time a character set of r characters is compared.
The number of verifications will continue to decrease.

Furthermore, when comparing the character set position information extracted from the search file, if the character sets that appear in all the keywords in the search input are checked in order of frequency, the search target will be further narrowed down and the number of matching matches will be reduced. It can be further reduced.

When character set position information matching the search input is found in this manner, the record corresponding to the record identification code corresponding to the search input is extracted and output to the searcher as a search result.

[0026]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Examples of the present invention will be described below with reference to the drawings.

FIG. 1 shows the configuration of an information retrieval processing device according to an embodiment of the present invention.

The information search processing device of this embodiment includes a CPU 1 that performs various arithmetic processing or judgment processing, programs for search processing, search file creation, etc., search files created or for performing search processing, search input, etc. a memory 2 for storing information, a keyboard 4, an input/output section 3 for connecting a display 5, an external storage device control section 6 for connecting an external storage device 7 for storing various information, a CPU 1, a memory 2, an input/output section 3, an external Common bus 8 connecting storage device control unit 6
Equipped with.

The information search process in this embodiment involves extracting constituent characters of each keyword one by one from the first character of the keyword string for a keyword string created from multi-keywords possessed by records to be searched for the search process. , a search file creation process that creates a character set with a total of two characters, that character and the next character, stores these character set types for each character set type, and creates a search file, and matching with the search file. The search process is divided into two: search processing, which extracts records with keywords that match the search input;

First, the search file creation process will be explained.

[0031] This search file creation process can be roughly divided into: ■ Securing a search file area; ■ Adding character set position information to each keyword constituent character set; ■ Adding character set position information grouped by character set type. It can be divided into three types: storage in files. Each of these processes will be explained.

① Securing search file area As shown in FIG. 2, the search file is composed of a character set group arranged in the order of characters listed in the JIS code table. As shown in FIG. 3, each character set group is composed of a character string of two characters in the order of characters listed in the JIS code table, with the listed characters being the first characters. Then, extract one character from the first character of the keyword of the record to be registered, and use that character and the next total of 2 characters.
Create character sets with characters and count the frequency of appearance for each type of character set. As a result, the number of character set position information registered in each character set type group constituting the search file can be known, so it is possible to secure an area for the search file composed of all character set type groups. At the same time, from the number of character set position information registered in each character set type group, the starting addresses of character set type groups that are successively stored in the search file can also be determined. The character set group address table shown in FIG. 4 is an arrangement of the leading addresses of the character set type groups in the order of description in the character set group list of FIG. 3.

■ Adding character set position information to each keyword constituent character set The character set position information described here is added to the keyword string created by arranging each keyword of a record in the keyword attribute area corresponding to the keyword attribute number. ,
The record number that indicates the registration order of records to which all character sets that make up each keyword belong, the character set position number that indicates the position of the character set in the keyword as the position of the first character of the character set, and the character set position number that indicates the position of the first character of the character set in the keyword. It is created with a keyword attribute number indicating the logical type.

First, record numbers will be explained. For example, in a general book search system, books are searched using keywords such as book title, author name, publisher name, year of publication, and abstract. At this time, the record includes the book name, author name, publisher name, publication year,
The search target consists of abstract keywords, and a record number is assigned by assigning a number in ascending order from 1 to the order in which this record is registered.

Next, keyword attribute numbers will be explained. Generally, a searcher inputs a keyword that is easy to identify the book to be searched, or inputs a stored keyword. For this reason, the book search system adds keyword attributes to each keyword, for example, book title, author name, publisher name, publication year, and abstract, and creates a logical association between the search input and the book search system's keywords. . Here, "1" is added to the book name as the keyword attribute number.
, "2" for the author name, "3" for the publisher name, "4" for the year of publication
, give the abstract a "5".

Next, character set position numbers will be explained. For each keyword, extract one character from the beginning of the keyword, create a character set with a total of two characters (that character and the next character), and assign numbers in ascending order of creation, such as 1, 2, 3, etc. Set position number. A special symbol EM (end mark) indicating the end of the keyword is added to the last character of the keyword, and the special symbol EM (end mark) is connected to the EM symbol to form a character set, and a character set position number is assigned. Note that an unused code from the JIS code table is assigned to the EM symbol. Here, the JIS code A3BA immediately after the number 9 is assigned.

Next, the keyword string will be explained. In order to realize a partial match search for keywords possessed by a record by matching a character set string with a search input character set string,
All the keywords of the record are concatenated to form a character string, and this is used as a keyword string. That is, each keyword is arranged in a fixed-length keyword attribute area corresponding to a keyword attribute number to create a keyword string. As a result, the attribute of the keyword to which the character set belongs can be determined from the character position in the keyword string. Note that, following each keyword attribute area, END symbols indicating the division of keyword attribute areas are arranged in the keyword string. This EN
The D symbol is assigned an unused code in the JIS code table. Here, the JIS code A3BB immediately after the EM symbol is assigned.

Then, for this keyword string, character set position information is created by converting all character sets constituting the keyword into codes consisting of integers from the record number, keyword attribute number, and character set position number. This character set position information is an integer code given by the following equation (1). Character set position information code = record number x n + (Pa
-1) +p...(1) n: Number of characters in the keyword string Pa: First character position in the keyword string in the keyword attribute area of keyword attribute number a p: Character set position number

For example, the size of the keyword attribute area of the keyword string is book name = 64 bytes 32 characters, author name = 3
2 bytes 16 characters, Publisher = 64 bytes 32 characters, Publication year = 8 bytes 4 characters, Abstract = 400 Bytes 200 In a book search system with 200 characters, the record number is 100.
The record is ``Book name = Structure of correspondence'', ``Author name = Ichiro Tanaka'', ``Publisher = Taiheiyo Publishing'', ``Year of publication =
1990'' and ``Abstract=Even for first-time users'', the keyword string is as shown in FIG. At this time, since the keyword string is 578 bytes (289 characters), the character set position information for each character set is configured as shown in FIG.

[0040] If character set position information is constituted by each 4-byte code in this way, it is possible to handle 232÷289≈14.86 million keyword strings each having 289 characters.

(1) Registration of character position information in the search file Next, the character set position information assigned to each character set is registered in the search file. As described above, character set type groups are stored in the search file in the order they are listed in the JIS code table. Then, character set position information for each character set is registered in each character set type group. Registration of this character set position information is performed by storing the character set position information at the beginning of each unstored area of the character set type group. Therefore, if record numbers are assigned in the order of registration, character set position information will be registered in ascending numerical order within the character set type group.

FIG. 7 shows an example in which the character set position information for the book name "Structure of Correspondence" mentioned above is registered in the search file. At this time, the character set position information within each group is stored in ascending order. This file size is 4 bytes of character set position information.

[Math 1] become.

[0043]Additional registration of character set position information is as follows:
This is done by adding a new code to the beginning of the unstored area of the group corresponding to each character set of each keyword included in the additional record. Further, deletion is performed by changing the corresponding character set position information in the group corresponding to each character set of each keyword included in the deletion record to a special symbol. This allows additional registration and deletion to be performed in a short time.

As mentioned above, the character set position information stored for each character set type group in this search file can be retrieved using the first address of each character set group in the character set group address table in FIG. 4 as a directory. .

The flow of the above search file creation process is shown in FIGS. 8a and 8b.

That is, a character set field address table is created by counting the frequency of appearance of character set types (S11, 12).
, secures an area for the search file (S13). Next, initialize the record registration rank counter to k = 1, set the record number to "1", and set the number of keyword string characters to n = 289.
The first character position of the keyword attribute area is P1 = 1 for the book name, P2 = 34 for the author name, P3 = 51 for the publisher name, P4 = 84 for the publication year,
P5=89 is set for the abstract (S14). Then, the first record is taken out (S15). This is the pre-processing for registration. From here, the registration process begins for each record. First, the keyword attribute number is set to a=1 (
S16), the keyword with keyword attribute number a is extracted from the record (S17). Furthermore, the number of constituent characters of the keyword is set to m, the character set position number is set to p=1, and the first character position of the keyword attribute area corresponding to keyword attribute number a is set to Pa (S18). Next, in order from the first character of the retrieved keyword, character set position information corresponding to character set position number p is calculated using the equation (2) of D=k×n+(Pa −1)+p.
) (S19).

Then, the character set field directory (character set field first address) indicating the character set field of the search file in which the character set type group of the character set at character set position number p is stored is retrieved from the character set field address table. The character set position information is retrieved (S20) and stored at the beginning of the unstored line in the character set column of the search file indicated by the character set column directory (S21). And p=
When p+1 and m=m-1 are set and all character sets in the keyword have been processed, the keyword attribute number is +1 with a=a+1 and the process moves to the next keyword (S24,
S25). Furthermore, when all the keywords included in the record are processed, the record registration ranking counter is incremented by 1 at k=k+1 and the process moves on to the next record (S26, S27).
, S28). When all records have been processed, the registration process ends (S26).

Next, search processing using the search file created in this manner will be explained.

In this embodiment, character set position information extracted from a search file is used to match keywords containing the same character string as the search input string, and after confirming that it has the same attributes as the search input, , an example of searching for records common to all search input character strings will be explained.

First, the search process can be roughly divided into the following components. ■ Break down the search input character string into character sets of two characters starting from the first character, and create a search input character set string. ■Calculate the first address of a character set group in the character set group address table that corresponds to each character set in the search input character set string. ■Reorder the search input character set string in order of character sets that appear least frequently. - Extract the corresponding character set type group from the search file in order from the beginning of the rearranged character set string, and extract the combination of character set position information that can constitute the search input character set string from the character set position information stored there. . (2) Extract character set position information having the same attribute as the search input from the extracted character set position information and make it a match. After repeating ■■ to ■ for the number of search inputs, the record number common to all search input character strings is output as the search result.

Next, each process will be specifically explained.

■ Creating a search input character set string In order to match the character set stored in the search file, the search input character string is broken down into a character set of two characters starting from the first character, and the search input character set string is created. shall be. Note that when a search input character string is decomposed into a character set of two characters from the first character, one character may remain at the end, making it impossible to create a character set of two characters. At this time, the last 1
A character set of two characters is created by extracting the missing character from the character set immediately before the character and concatenating it to the front of the last character.

[0053] Calculation of the character set group start address in the character set group address table corresponding to each search input character set Similarly to the time of creating the search file, calculate the character set description rank in Figure 3 for each search input character set. This is then used as the address pointer for each search input character set in the character set group address table.

■ Rearranging the order of appearance frequency Then, by referring to the character set group start address in the character set group address table that indicates the start address of each character set type group in the search file, the appearance frequency of each search input character set is determined. Search and rearrange the search input character set strings in descending order of appearance frequency among all keywords. As mentioned above, the first address in the character set group address table indicates the first address of each character set type group stored in the search file, and if you take the difference from the first address of the next character set group, , from the number of character set position information stored in each character set type group, the appearance frequency of the character set type in all keywords can be determined.

[0055] This is because the number of times of matching between the character set position information of each character set stored in the search file can be extremely reduced by performing matching from character sets that appear less frequently. In other words, when checking the continuity of each character set by comparing character set position information, in order to check the record number, character set position number, and keyword attribute number in the character set position information in two character set type groups, The smaller the number of character set position information stored in the two character set type groups, the fewer times the matching can be performed. therefore,
When performing character set position information verification, performing verification from character sets that appear less frequently reduces the number of times of verification. In particular, as the number of search input characters increases, the rate of inclusion of character sets with low frequency of appearance increases, so the reduction effect is large.

① Character string comparison The character set position information stored in each character set type group is extracted by referring to the character set group address table from the character set with a low appearance frequency. Then, based on the retrieved character set position information, the characters in the search input string that have the same record number and the difference in character set position number between each character set type group are searched in descending order of the character set type groups with the lowest appearance frequency. Extracts a combination of character set position information that is equal to the first character position difference of the set.

[0057] This character set position information is collated by setting i to the character set position number of a character set that appears less frequently in all keywords in the search input character set string, and j to the character set position number of a character set that appears frequently. Then, in the following equation (3), (character set position information of the character set with character set position number i) - (character set position information of the character set with character set position number j) = i - j ... (3) It is sufficient to extract combinations of character set position information.

In this process of collating character set position information between character set type groups, the character set position information of the character set type group that appears less frequently is compared with the character set position information of the character set type group that appears more frequently. Match consecutive characters by taking the difference between them.

Assuming that the arbitrary character sets in the search input character string are AB and CD, in order to extract the corresponding character set position information, assume that the character set position number difference between AB and CD is L, and extract the character set position information of group AB. Character set position information A
x, the character set position information of group CD is Cy.If Ax +L>Cy, delete Cy.If Ax +L<Cy, delete Cx.If Ax +L=Cy, delete both Ax and Cy as a match.Discontinuous characters The number of times of matching is reduced by deleting the set position information from the matching target.

For example, the character set position information of group AB is 571, 1140, 10230, 11365, 1
The character set position information of 1928, 13069 group CD is 569, 1706, 10231, 10235, 1
When the 1362 character set position number difference L=2, the number of times the two groups are compared is only seven times in total, and there is no need to compare all the character set position information within the group.

③ Matching of Keyword Attribute Number The keyword attribute is matched with respect to the character set position number of the character set position information obtained from the character string matching. That is, if the character set position number is 1 to 32, the keyword attribute of the character set position information is the book name, and if the character set position number is 34 to 49, the keyword attribute of the character set position information is the author name, and the character set position information is the author name. If the position number is 51 to 82, the keyword attribute of the character set position information is the publisher name, and if the character set position number is 84 to 87, the keyword attribute of the character set position information is publication year, and the character set position number is the publisher name. 89 to 288, it can be seen that the keyword attribute of the character set position information is abstract. Therefore, only the character set position information that is the same as the attribute specified at the time of search input is extracted from the character set position information obtained by character set string matching.

[0062] ■ Extracting record numbers Repeat ■ to ■ for the number of search inputs, and extract the record number common to all search input strings among the character set position information corresponding to each search input string obtained. . Note that in this embodiment, the case where there is one or more search inputs has been described. In addition, when there are multiple search inputs, we explained as an example of performing a logical AND operation between each search input, but in the case of multiple search inputs that involve logical operations other than a logical AND operation, the records that remain as a matching result A specified logical operation is performed by associating the number with each search input, and a record number that satisfies the logic of the search input is output as a search result.

[0063] When specifying one search input with multiple character strings, which is often the case when targeting items with a large number of characters such as abstracts, after completing the keyword attribute matching of the first character string, For the second and subsequent strings,
The character set position information with the record number and keyword attribute number obtained so far is extracted from the character set type group of the first collation character set of that character string, and the obtained character set position information is used at the beginning of character set string collation. Performs matching processing on other character sets within the same character string as a character set type group of characters.

The above operations ① to ② will be explained using specific examples.

Assume that a book name is specified as a search target, and ``correspondence'' is specified as a search input character string. In this case, the attribute number of the keyword attribute of the book name is "1"
shall be.

Since the search input is "correspondence," the search input character set is "correspondence,""document," and "calligraphy." Assuming that the frequency of occurrence of all keywords is in the order of "communication"<"document"<"calligraphy" and matching is performed in this order, first, the character set extracted from the character set group field of "communication" in the search file. Between the location information and the character set location information extracted from the character set group field of "Document", "Tsu" and "Tsu" in the search input "Correspondence" are
Since the character positions with "Sentence" are "1" and "3" respectively, if character set position information with a character set position difference of "-2" is extracted, the characters in "Communication" in the search file in Figure 7 will be extracted. "28901" in the set position information and "28" in "Document"
903'' can be extracted as a combination of continuous character set position information. Next, the character set position difference between the character set position information that remained as a matching result in "Communication" and the character set position information extracted from the character set group field of "Sho" in the search file is "
-3'' character set position information, the character set position information in ``Communication'' of the search file in Figure 7 is ``2''.
8901” and the character set position information in “Sho no” “2”
8904'' can be extracted as a combination of continuous character set position information. Furthermore, since the keyword attribute is "book name", "289
01'', ``28903'', and ``28904'' can be extracted.

Since the number of characters in the keyword string is "289", the character set position number is 28901÷289.
=100 From the remainder 1, we know that it is 1. It can also be seen that the record number to which this character string belongs is 100.

This search processing operation is shown in flowcharts in FIGS. 9a and 9b.

In other words, the number of characters in the keyword string is n=28.
9, the character position range Pa of the keyword attribute area is P1 = 1 to 32 for the book name, and P2 = 34 to 4 for the author name.
9, Publisher name is P3 = 51-82, Publication year is P4 = 8
4 to 87, the abstract is set to P5 = 89 to 288, the keyword attribute number is set to a = 1 (S31), and if there is a search input character string with keyword attribute number a, it is extracted (S32, S33). From here on, the search input string will be matched. Therefore, a search input character set string is created by taking the search input and dividing it into character sets of two characters from the beginning of the search input character string, and setting the number of character sets minus 1 as the number of matching times q (S34, S35). The search input character set string is rearranged in descending order of appearance frequency among all keywords (S36). Then, extract the character set position information stored in the character set type group corresponding to each rearranged character set from the search file (S3
7). Next, the character set position number of the character set that appears less frequently in all keywords in the search input character set string is i
, set the character set position number of the character set that occurs infrequently to j
When, between two character set type groups, (character set position information of character set with character set position number i)
-(Character set position information of the character set with character set position number j)=i−j is extracted (S38). Similar processing is performed for the remaining character sets in the search input character set string (S39, S
40), character position range Pa whose character set position number is keyword attribute number a from the remaining character set position information
Extract only the record number within. To extract the character set position number from the character set position information, use the following formula (
4) is used. (Character set position information) ÷ (Number of keyword string characters) = Record number Remaining character set position number... (4) In the processing up to this point, you have the search input string as a character string, and you also have the same keyword as the attribute entered in the search. The record number that the user has is known (S41). The same process is performed up to the abstract, and a record number having the same keyword as the attribute entered in the search is extracted (S42, S43). When all the search input strings are compared, the record numbers common to all the search input strings are output as search results for the remaining record numbers (S44).

[0070] In the above embodiment, an example of search processing for a Japanese document was explained, but as another example, for a document in a language with few character types, such as English characters, kana characters and kanji characters are searched from the JIS code table. By creating a search file, character set group list, and character set group address table using only character types excluding A search file for searching can be constructed, and the search can be easily performed. In this case, for English characters (including symbols), instead of using JIS codes, for example, ISO codes,
It is preferable to create a character set group address table using ASCII codes, and to make the character set consist of at least three characters.

[0071] When high-speed multi-keyword search is required, increasing the number of characters constituting a character set will cause the character set to appear less frequently, and the character set position information stored in each character set type group will decrease. Therefore, speeding up can be easily achieved.

Further, in the case of a Japanese document, a search file may be created in which only consecutive katakana or hiragana characters are set as a character set, and kanji are set in units of characters. Kanji has many character types and their appearance frequency is low, so
It is also possible to speed up the search by using a character type group for each character.

[0073]

As explained above, the present invention allows the record identification code, character set position order code, and keyword type to which the character set belongs to be determined for each character set type of the constituent characters of the keyword included in the record to be searched. Create a search file that stores character set position information consisting of the keyword attribute number shown, extract the character set position information for each character set type that makes up the search input character string, and use this search file to search for characters that include the search input. Added column search.

[0074] Therefore, (1) the number of character string comparisons for search processing can be reduced, making it possible to perform a high-speed search; (2) search processing can be performed by focusing on the character set and character position. (3) High-speed searches can be achieved using only software without the need for specialized hardware, so partial match searches using multiple keywords can be performed efficiently with general-purpose information processing equipment. (4) When used in a database system, there is no need to create a partial match search string that can be used to create a search file, and automatic search is performed from the keywords of the records to be searched. Because search files can be created automatically, it is possible to construct a database system economically and efficiently. Since the frequency of occurrence of each character set is low because the frequency of occurrence of the column is low, high-speed search is possible by creating a search file that stores character set position information in the character set type group that makes up the character string. It has an excellent effect.

[Brief explanation of drawings]

FIG. 1 shows a configuration example of an information search processing device used in an embodiment of the present invention.

FIG. 2: Example table of search files in the embodiment.

[Figure 3] List of character set groups in the example.

FIG. 4: Character set group address table of the embodiment.

[Figure 5] Examples of keywords in the example.

FIG. 6 is an example of creating character set position information according to the embodiment.

[Fig. 7] Example of search file of the embodiment.

FIGS. 8a and 8b are flowcharts illustrating search file creation processing procedures in the embodiment.

FIGS. 9a and 9b are flowcharts illustrating the search processing procedure of the embodiment.

[Explanation of symbols]

1...CPU 2...Memory 3...Input/output section 4...Keyboard 5...Display 6...External storage device control unit 7...External storage device 8...Common bus

Claims (7)

[Claims] 1. Record identification code assigning means for assigning an ascending code to each record to be searched, and keyword attribute code assigning means for assigning an attribute code indicating a logical classification of the keyword to each keyword included in the record. Then, this keyword is broken down into individual characters to create a character set consisting of that character and the next r characters in total (where r is a natural number greater than or equal to 2), and in this character set, the first character set in the keyword is added to the character set. Character set position order code adding means for adding a character set position order code indicating character position;
Means for creating character set position information consisting of the record identification code, keyword attribute code and character set position order code, and storing this character set position information in an area for each character set type to create a search file. Equipped with an information search processing method. 2. The character set position information includes record identification code x n + 2. The information search processing method according to claim 1, wherein (Pa -1)+character set position order code n: number of characters in keyword string Pa: given as a numerical code representing the first character position in the keyword string in the keyword attribute area of keyword attribute code a. [Claim 3] Regarding the keyword string of the record to be searched, record identification codes assigned to each record in ascending order for each character constituting each keyword, and keyword attribute codes indicating the logical classification of the keywords included in this record. and a character set position order code indicating the first character position of that character set in the keyword. A means for configuring a search input character set string decomposed into character sets of r characters, and extracting character set position information of the same character set as the decomposed character set from the search file, and a character set position of each extracted character set. A character set position where the record identification code and keyword attribute code are common among the information, the character set position order code is equal to the first character position difference of the corresponding character set of the search input string, and the keyword attribute code is equal to the search input. An information search processing method comprising means for extracting a combination of information, and means for outputting a record identification code corresponding to a search input as a search result based on the extracted combination of character set position information. 4. The extraction of combinations of character set position information that can form the same character set string as the search input character set string is performed in descending order of the frequency of appearance in all keywords of the search input character set string. information retrieval processing method. 5. Extraction of combinations of character set position information that can constitute the same character set string as the search input character set string is performed by extracting character set position order codes of character sets that appear less frequently in all keywords of the search input character set string by i. ,
When the character set positional order code of a frequently occurring character set is j, (character set positional information of the character set with character set positional order code i) - (character set positional information of the character set with character set positional order code j) )=i−j The information search processing method according to claim 3 or 4, wherein a combination of character position information matching the following is extracted. [Claim 6] When the keyword is a European character string that includes a symbol, a search file for only character set types of European characters that include symbols is used in a character set of at least three characters and symbols. The information retrieval processing method described in [Claim 7] When the keyword includes a kanji character, a search file consisting of a character set type of at least two characters for kana characters is used.
The information retrieval processing method described in any of the above.