JP3046221B2 - Information retrieval device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information retrieval apparatus for retrieving data including an input keyword from data registered in a database, and more particularly, to accurately retrieve data from a database such as a table. It is made possible.

[0002]

2. Description of the Related Art In recent years, in information retrieval, when a user enters a keyword as a keyword, a full-text search is performed on data registered in a database, and a system for automatically retrieving data including a character string of the keyword has attracted attention. Are gathering. In particular, products such as relational database management systems (RDBMS) are being commercialized.
Applications in various fields are spreading.

In the relational database, for example, as shown in FIG. 22, when the record of record number 1 is a record relating to Yokohama in Kanagawa prefecture and the record of record number 2 is a record relating to Kamakura in Kanagawa prefecture,
A plurality of items representing record attributes (prefecture name, city name, ‥) are provided as a data structure, and data of each attribute is stored in each item in correspondence with the record number to form a table.

When a record of, for example, "Kamakura" is searched from the database of this table, first, as shown in FIG. 23, the operator inputs an item (item 2) of the table to be searched into the information search device. And the keyword "Kamakura" to be searched. In response to this, the information search device reads data relating to all records in the table of FIG. 22, sets item 2 designated as a search target as a target to be checked, and checks a keyword against each data of item 2. I do. When matching data is detected by this collation, a record number (number 2) corresponding to the data is output.

[0005] As described above, the conventional information retrieval apparatus retrieves a target record by using the items and keywords input by the operator as clues.

[0006]

However, in the conventional information retrieval, for example, if one record is a record relating to "Chinatown" of "Yokohama" and the other is a record relating to "Yamashita Park" of "Yokohama" In order to distinguish these records, it is necessary to add a new item related to “Famous Places” and store data of “Chinatown” and “Yamashita Park” in correspondence with each record. There is a problem that it increases. Also, in this case, when searching for a record related to Chinatown in Yokohama, it is necessary to specify a plurality of items, one for searching for "Yokohama" and the other for searching for "Chinatown". Since it is necessary to know in advance whether necessary data is stored, there is a problem that a search operation becomes complicated and a search process takes time.

In such a case, there is also a method of listing a plurality of data in one item (in the above example, "Yokohama Chinatown").
"Yokohama Yamashita Park"), but at this time, "Yokohama"
Since "Chinatown" and "Yamashita Park" do not exist as independent data, for example, when it is desired to search for a record having a single data of "Yokohama", these records are omitted from the search. At this time, if a search is performed that picks up all data that includes the word “Yokohama”, these records can be prevented from being omitted. However, in such a search, data contrary to the searcher's intention (search garbage) ) May be detected in large quantities.

[0008] When a plurality of data are described in one item as described above, a character string of the data becomes long, and it takes time to sequentially collate input keywords from the beginning of the data.

[0009] In the search using this keyword, for example, data having a "horizontal" in front of it, such as "horizontal XX" or "XX horizontal", is searched (hereinafter, referred to as "front match search"). It is possible to search for data with "horizontal" at the back (hereinafter "rearrangement search") or to search for data with characters specified both forward and rearward (hereinafter referred to as "perfect match search"). However, these search times also become longer.

The present invention solves such a conventional problem. A plurality of data can be independently registered in one item provided as a constituent element of a table. It is an object of the present invention to provide an information retrieval apparatus capable of performing high-speed collation between data and input keywords.

[0011]

Therefore, according to the present invention, there is provided an information retrieval apparatus for retrieving data including an input keyword from data in an item of record data having a plurality of items. When composed of a plurality of data, a delimiter providing means for providing a delimiter code between the beginning and end of the column connecting these data and the data, and a character for creating a character chain consisting of a set of two adjacent characters in this column A chain pattern generating means and character chain information including the number of appearances in each character data constituting a character chain, a record number, and a data number are created, stored in a memory in association with the character chain, and a search file is created. Search file creating means, keyword delimiter assigning means for assigning a delimiter code to the input keyword,
Keyword character chain generation means for generating a character chain of keywords from keywords to which a delimiter code has been assigned; and character chain information corresponding to the character chains of the keywords are read from a search file and collated to search for data including the keyword. Collation means is provided.

Further, the matching means is provided with priority matching means for giving priority to matching of character chain information corresponding to a character chain of a keyword including a delimiter code.

[0013] The character chain pattern generating means creates a character chain by including the delimiter code in the character chain.

At this time, the search file creating means includes:
As the number of occurrences of the delimiter code in the character chain information,
Set a specific number.

Further, the character chain pattern generating means creates a character chain without including the delimiter code, and the search file creating means converts the character chain information of the character chain adjacent to the delimiter code out of the character chains. Created separately from character chain information.

[0016] Further, the search file creating means stores the character chain information of the character chain adjacent to the delimiter code in the memory for each group.

Further, the delimiter assigning means assigns a plurality of types of delimiter codes indicating the hierarchy to the preceding column, and the keyword delimiter assigning means assigns the input keyword with the delimiter code indicating the hierarchy.

Further, the delimiter assigning means assigns a plurality of types of delimiter codes indicating the order to the preceding column, and the keyword delimiter assigning means assigns a delimiter code indicating the order to the input keyword.

[0019]

Therefore, it can be checked whether or not continuous characters corresponding to the character string of the keyword exist in the same data by comparing the character chain information read out in correspondence with the character chain of the keyword. Even when a plurality of such data exist in one item, each data acts as an independent data because each data is separated by a delimiter code. Therefore, a plurality of data can be repeatedly registered for one item which is a component of the record data, and data such as a character string and a pattern including the input keyword can be repeatedly searched in the item. is there.

Further, a delimiter code is added before or after the keyword, and character chain information corresponding to a character chain including the delimiter code is preferentially collated.
It is possible to perform a prefix search and a tail match search at high speed.

When a character chain adjacent to the delimiter code is treated as the head or end of data without including the delimiter code in the character chain, the number of character chain information can be reduced, and the number of times of collation can be reduced. Can be reduced. Also, by storing the character chain information of the character chain adjacent to the delimiter code in the memory for each group, reading from the memory becomes easy. In addition, the reading order and collation order of the character chain information can be matched in accordance with the forward match search, the backward match search, and the perfect match search, and these searches can be performed at high speed.

Further, for data having a hierarchical structure or an order structure, a delimiter code indicating the hierarchy or order is assigned to form a search file, and a keyword to be input is also assigned a separator code for designating the hierarchy or order. By doing so, a search having a hierarchical relationship or an order relationship can be realized.

[0023]

Embodiment (First Embodiment) In the information retrieval apparatus of the embodiment, as shown in FIG. 6, when there are three data items "Ah, Ah, Ah," in a record item, In place of the delimiter "," in the data, a delimiter code (@W) is added, and the data is converted into data "@W," AW, AW, AW, W ". Next, the number of character appearances for each data is obtained. This character appearance count indicates the number of times each character appears in a character of the same type.
The result is as shown in FIG. Next, a character chain pattern consisting of a set of two adjacent characters while moving one character at a time from the beginning of "@W aa ae @ W aa ai @ W aa aa @ W", that is, (@W, a) (a , I) (i, u) (u,
A) (A, E) (E, @W) (@W, A) (A, I)
(A, A) (A, E) (E, @W) (@W, A) (A,
(), (A, a) (a, s) (s, ∧W) are created, the number of appearances of each character in the character chain pattern, a data number (a number indicating the order of three data), Create character chain information by combining with the record number.

For example, the character chain information of (a, i) which is the second character chain pattern is such that the number of appearances of the character "a" is 1, the number of occurrences of the character "i" is 1, the data number is 1, and Since the record number is 1, it becomes (1,1,1,1). Further, the character chain information of the last character chain pattern (（, ∧W) is (1, m, 3, 1). Here, the number of appearances of the character ∧W is represented by m. A search file (FIG. 6 (4)) is created by combining the character chain patterns thus obtained and the character chain information.

On the other hand, if a keyword "rear toe" is inputted as shown in FIG. 10 at the time of retrieval, a delimiter code is added to the keyword and "@W rear toe"
W ”, and its character chain pattern ($ W, a)
Character chain information corresponding to (A, A) (A, A) (A, S) (S, $ W) is obtained from the search file (FIG. 10).
(4)).

Next, it is checked from the character chain information as to whether or not a continuous character string "rear foot" exists in the data 1 to 3 as follows. If the character string exists in the data 1 to 3, the data number and the record number of the character chain information in each of the character chain patterns should match, and each character chain pattern , The value of the second element of the character chain information in the next character chain pattern should match the value of the first element of the character chain information in the next character chain pattern. When character chain information satisfying such conditions can be detected in the character chain information obtained from the search file, the record number is output as a detection result.

As shown in FIG. 1, the information retrieval apparatus which performs such an operation includes a record data storage section 2 storing record data composed of a plurality of items, and a record reading section 1 for reading the stored record data. And the data of each item is cut out from the read record data, the order of the data is counted,
A delimiter adding unit 3 for adding a delimiter code at the end of the
While acquiring the order of the data from the delimiter addition unit 3,
A number adding unit 9 for acquiring a record number from the record data storage unit 2; a character chain pattern generating unit 7 for calculating the number of character appearances in the data of each item to which a delimiter code is added and creating a character chain pattern; A segment determining unit 8 for creating a character chain pattern including a code; a character chain based on the character chain pattern obtained from the character chain pattern generating unit 7 and the segment determining unit 8 and the data number and record number obtained from the number adding unit 9 A search file generation unit 10 for creating information; a search file storage unit 11 for storing character chain patterns, character chain information, and data of character types for delimiter codes; input keywords and partial matches, forward matches, backward matches A keyword reading unit 12 that reads each of the search types, that is, a match and an exact match, is separated before and after the keyword according to the search type. Keyword addition unit 13 for adding a code, a keyword character chain generation unit 17 for generating a character chain pattern from keywords to which a delimiter code has been added, and searching for character chain information corresponding to the character chain pattern generated from the keyword It includes a collating unit 18 for extracting collating data from the file 11 to perform collation processing and extracting record data matching the keyword, and a search result display unit 19 for displaying search results, and further includes a character including a delimiter code in a second element. The priority instructing unit 6 for instructing the retrieval file storage unit 11 to store the character chain information in the retrieval file storage unit 11 at a position where it can be preferentially extracted, and the character chain information of the character chain including the delimiter code in the second element are preferentially compared. The priority collating unit 16 instructs the user to perform a delimiter command for the last two characters of each data included in one item. Group addition unit 4 for adding a keyword, a keyword group separation addition unit 14 for adding a separation code to the character type combination of the last two characters of the input keyword, and a plurality of items included in the record data item. When the data has a hierarchical structure, a hierarchical / sequential delimiter 5 for adding a delimiter code according to the hierarchical structure to the end of each data, and a keyword hierarchical delimiter for adding a delimiter code indicating the hierarchy to the input keyword. An additional unit 15 is provided.

In the first embodiment, the operation of generating a search file and the operation of searching a keyword in this information search apparatus will be described in detail. In this operation, the priority instructing unit 6, the priority collating unit 16, the group-specific delimiter adding unit 4, the keyword-group-specific delimiter adding unit 14, the hierarchy / order-specific delimiter adding unit 5, and the keyword-level delimiter adding in the configuration of FIG. Part 15 is not used.

As shown in FIG. 3, the structure of the code data storage unit 2 is composed of record numbers, items, and the like, and each item stores a plurality of data (da).
ta1, data2, ‥). In this embodiment, it is assumed that data exists in item 2. The procedure up to generating the search file from the record data stored in the code data storage unit 2 is performed according to the processing flow of FIG.

Step 1: The record reading section 1 reads the record data from the record data storage section 2 and sets the first record data as record data to be processed (L = 1). The item to be processed (the k-th item here) is specified. As shown in FIG. 6 (1), it is assumed that three data items "Ah, Ah, Ah, and Ahashi" are separated by "," in the designated item.

Step 3: The delimiter adding unit 3 adds a delimiter code (@W) to the beginning and end of the data, and changes a delimiter "," between data into a delimiter code.
FIG. 6B shows data to which a delimiter code is added.

Step 4: Character chain pattern generator 7
Sets the data number h of the data to be processed of the item k to 1; Step 5: sets the number of appearances for the character type of the delimiter code in the data of the data number 1 to a specific number (here, m) Step 6: For each character of this data, the number of appearances for each character type is calculated. FIG. 6 (3) shows first to third
Obtained when calculating the number of occurrences of each character type for all data, shows the relationship between data different appearance count and delimiter code (∧W).

Step 7: Character chain pattern generator 7
Creates a character chain (F (a), F (b)) in the data of data number 1. At this time, the delimiter determination unit 8 creates a character chain pattern including the delimiter code.

Step 8: The search file generator 10 obtains the character chain pattern and the number of appearances thereof from the character chain pattern generator 7 and the delimiter judging unit 8, and obtains the data number (h ) And the record number (L) to create character chain information and store it in the search file storage unit 11. This character chain information includes, as shown in FIG. 4, the number of appearances of two consecutive characters, a data number,
Consists of a record number.

Step 9: In this way, character chain information in all character chain patterns of the data with the data number h is created.

As shown in FIG. 5, the search file storage unit 11 is composed of a type of character chain information, a character chain pattern, and a list of character chain information. It distinguishes whether it is included and if it is, whether it is the front or back of two consecutive characters. Step 10: When F (a) = m, this character chain The character chain information is stored as character chain information with forward delimiter code. Step 11: When F (b) = m, the character chain information of this character chain is stored as character chain information with rear delimiter code.

Step 12: Next, the processing target is item k
In order to move to the next data in, the data number h is incremented. Step 13: The procedure from step 5 is repeated until the value of h reaches the maximum data number in the item k.

FIG. 6 (4) shows a search file generated from the data of FIG. 6 (1) and stored in the search file storage unit 11.

In step 13, when the value of h exceeds the maximum data number in the item k, step 14: increment the record number L to move the processing target to the next record data, and step 15: Steps 2 to 15 are repeated until the processing for all record data is completed.

Next, a description will be given of a processing procedure for searching for data that matches an input keyword using a search file, with reference to the processing flow of FIG. FIG. 8 shows the flow of data from the reading of the keyword to the search at this time. FIG. 8 shows data read by the keyword reading unit 12.
The pattern 1 “After?” And the pattern 2 “Aie” in (1) will be described as examples. "After?"
This means that a character string that includes "after" at the beginning is checked, and if there is no "?" Like "Oh," it means that a character string that completely matches the keyword is checked.

Step 1: The keyword reading section 12 reads the input keywords of pattern 1 and pattern 2, and sets the item k input by the operator as a designated item.

Step 3: Keyword delimiter 13
Is the search type (partial match, start match, end match, perfect match) according to the presence or absence of "?" In the keyword string
Step 4: Delimiter code (@W) is assigned to the keyword according to the search type. As shown in FIG. 8 (2), when the search type of the keyword is an exact match, a delimiter code is added to both ends of the character string, and when the keyword character string has "?" Add a delimiter code only to the end of.

Step 5: Keyword character chain generation unit 17
Creates a character chain pattern (a, b) for a keyword to which a delimiter code has been assigned. FIG. 8 (3) shows a character chain pattern corresponding to pattern 1 and pattern 2.

Step 6: Next, the collating unit 18 extracts character chain information corresponding to the character chain pattern from the search file storage unit 11 and performs collation. If the character type a is a delimiter code, Step 7: extract all the character chain information of the corresponding character chain pattern from the character chain information with the forward delimiter code of the search file. Step 8: If the character type b is the delimiter code Step 9: Extract all the character chain information of the corresponding character chain pattern from the character chain information with the rear delimiter code of the search file. Step 10: Otherwise, extract the character chain information without the delimiter code of the search file. All the character chain information of the corresponding character chain pattern is extracted from the inside.

Step 11: Next, the number of character chain sets of the keyword is checked, and if the number of sets is one, Step 12: all the extracted character chain information is used as the corresponding character chain information.

Step 14: If there are two or more sets, Step 15: The number of appearances of the second character of the first character chain and the second character chain occurrences of the first character matches the, and verifies whether or not the data number matches exists, repeat the collation replaced one after another character chain information until they match.

Step 16: When the corresponding character chain information exists in the collation, Step 17: the collation unit 18 extracts the corresponding record number and the number of records from the record data storage unit 2, and displays the retrieval result display unit. Display at 19.

FIG. 8 shows the data flow at the time of collation at this time.
This is shown in (4). In the case of <pattern 1> in FIG. 8D, since the collation between the first character chain and the second character chain (collation with (A) and (B)) does not match in the first collation, the second collation is performed. The collation with the next first character chain information is performed without changing the character chain information (second collation). Since the matching is repeated the third time, the matching unit 18 determines the item number 2, the data number 3, the record number 1 from the matched character chain information.
Is obtained from the record data storage unit 2. If all the character chain patterns do not match, the collation ends.

<Pattern 2> in FIG.
This is an example of matching against a character chain pattern.
(A) and (B), (C) and (B), (D) and (B), and (E) and (B) are shown the process of collating. FIG. 8 (5) shows the final search result.

As described above, in the information retrieval apparatus of the embodiment, a delimiter code is added to the beginning, interval, and end of data, a character chain pattern is formed including the delimiter code, and character chain information of each character chain pattern is collected. To form a search file. Then, the input keyword is decomposed into a character chain including a delimiter code, and character chain information corresponding to the character chain is collated to detect data including a character string of the keyword.

As described above, since the delimiter code is assigned to the data in the item, even when a plurality of data are included in one item, the data can be made to act as independent data in the keyword search. .
Therefore, it is possible to add a plurality of data in one item without causing any trouble in the search.

Also, at the time of retrieval, only the character chain information corresponding to the character chain of the keyword is read from the search file and collation is performed. Therefore, compared with the case where the keyword character string is collated with all the data character strings, the retrieval time is longer. Can be shortened.

Although the embodiment has been described with reference to the case where the data is a character pattern, the present invention can be applied to any pre-patterned data such as △. As the delimiter code, any code other than the patterned data can be used.

(Second Embodiment) In the second embodiment, a case will be described in which collation with an input keyword is performed from the end of a keyword character string in order to facilitate a search such as backward matching.

In this operation, in the information retrieval apparatus of FIG. 1, in addition to the parts used in the first embodiment, character chain information of a character chain including a delimiter code in the second character is stored in a preferential reading position. This is performed using a priority instructing unit 6 for instructing to perform collation and a priority collating unit 16 for instructing to preferentially collate the character chain information of the character chain including the delimiter code in the second character.

In this apparatus, the generation of the search file is performed in the same manner as in the first embodiment. However, regarding the character chain information of the character chain including the delimiter code in the second character, the search file generation unit 10 receiving the instruction of the priority instruction unit 6
Is stored in the search file storage unit 11 at a position where reading is possible preferentially.

On the other hand, a search for a match with the input keyword is performed according to the flowchart of FIG. FIG. 10 shows the flow of data from reading of a keyword to searching at this time.
Is shown in It is assumed that the input keyword is “rear toe” in FIG.

Steps 1 to 5: The keyword reading section 12 reads an input keyword and sets an item k input by the operator as a designated item. The keyword delimiter 13 determines a search type according to the character string of the keyword, and separates the keyword with a delimiter code (∧
W). The keyword character chain generation unit 17 generates a character chain pattern (a, b) of a keyword to which a delimiter code has been added. The procedure up to this point is exactly the same as in the first embodiment. FIG. 10 (2) shows the keyword to which the delimiter code is added, and FIG. 10 (3) shows the created character chain pattern.

Step 6: The collating unit 18 determines whether the character type b is a delimiter code in a character chain pattern (し, ∧W)
The character chain information of the character chain pattern (@W, a) in which the character type a is the delimiter code is extracted from the search file storage unit 11 with priority. Step 7: Next, the other character Extract the character chain information of the chain pattern from the search file.

Step 8: Next, the number of sets of character chains of the keyword is checked. Step 9: If the number of sets is one, Step 10: All the extracted character chain information is set as the corresponding character chain information.

[0061] Step 11: the two or more sets at a time, step 12: collates whether the number of occurrences and data number between the character chain information in the two sets of character chain consecutive matches, they match The character chain information is replaced one after another until the collation is repeated.

Step 13: When there is character chain information that matches in the collation, Step 14: The collation unit 18 extracts the corresponding record number and the number of records from the record data storage unit 2, and displays the retrieval result display unit Display at 19.

The procedure of steps 8 to 14 is the same as the procedure of steps 11 to 17 (FIG. 7) in the first embodiment. However, in the collation in step 12,
As shown in FIG. 10 (4), the character chain pattern (し, ∧W) in which the second character is a delimiter code is prioritized. First, the character chain information of (（, ∧W) and the preceding ( Ah)
To check whether the number of occurrences of the first character of (し, ∧W) matches the number of occurrences of the second character of (あ, ∧), and whether the data numbers match. . When they match, the character chain information of (A, S) and (A, A) is collated, and the collation is sequentially advanced after the keyword.

As described above, in the information retrieving apparatus of the second embodiment, the character chain having the delimiter code as the second character can be prioritized in the collation, so that the backward matching retrieval can be performed at high speed.

In the collation, it is possible to give priority to a character chain having a delimiter code as the first character, and in this case, a forward match search can be performed at high speed.

(Third Embodiment) In the third embodiment, the data following the delimiter code of the data to which the delimiter code is added,
A high-speed search is realized by treating characters as a character chain with a delimiter code.

In this operation, in the information retrieval apparatus shown in FIG. 1, in addition to the parts used in the first embodiment, a delimiter code is added to a combination of two character types at both ends of each data included in one item. This is executed using the group-specific delimiter adding unit 4 and the keyword-group-specific delimiter adding unit 14 that adds a delimiter code to a combination of two character types at both ends of the input keyword.

In this case, the search file is created as shown in FIG.
Is performed according to the processing flow of

Steps 1 to 7: When the item (k-th) to be processed of the record data is specified, the record reading section 1 reads the record data from the record data storage section 2, and the delimiter section 3 A delimiter code (@W) is added to the beginning and end of this data, and the delimiter "," between data is changed to a delimiter code. The character chain pattern generation unit 7 sets the data number h of the data to be processed of the item k to 1, sets the number of occurrences to the specific number m when the character type is the delimiter code, and sets the other characters of this data to For, calculate the number of appearances for each character type,
Next, a character chain (F (a), F (b)) is created.

The operation so far is substantially the same as that of the first embodiment, except that the character chain pattern generation unit 7
Is used to create character chains (F (a), F (b))
Based on the instruction of the group-specific delimiter adding unit 4, the delimiter code is treated as a group-specific delimiter code, and a character chain (F (a), F (b)) is created without including the delimiter code in the character type. Step 8: The delimiter determination unit 8 checks whether the first character or the second character of the character chain is adjacent to the delimiter code. One character is separated into a character chain with a forward delimiter code adjacent to the delimiter code, and the second character is separated into a character chain with a rear delimiter code adjacent to the delimiter code.

Step 11: The search file generation unit 10 separates and stores the character chain information of these separated character chain patterns in the search file storage unit 11. The configuration of the search file storage unit 11 is shown in FIG.

These operations are performed until the data number reaches the maximum in the item k, and is repeated for all record data.

FIG. 13 (1) shows the number of appearances according to the character type of the record data and the group-based character chain pattern connected to the delimiter code.
FIG. 13 shows a search file created from FIG. In this search file, (a, i), (a, and)
Are stored as a character chain pattern with a forward delimiter code, and are partitioned as a group of "A" by the first character. Similarly, (A, E) (A, S) is stored as a character chain pattern with a backward delimiter code, and is partitioned as a group of "A".

It should be noted that (a, i), (a,
), (A, E) (A, S) are not stored as character chain information without a delimiter code. That is, the character chain information to be collated is reduced accordingly, and the collation can be performed at high speed.

Next, FIG. 14 shows the flow of the retrieval process.
FIG. 15 shows the flow of data at the time of retrieval. Here, it is assumed that the input keyword is “rear toe”.

Steps 1 to 3: As in the first embodiment, the keyword reading section 12 reads a keyword, receives an input operation by the operator, sets the keyword in a specified item, and specifies a search type.

Step 4: Keyword Group Delimiter Addition Unit
14 adds a group-based delimiter to the keyword (FIG. 15)
(2)), the keyword character chain generation unit 17 divides the keyword into character patterns to create a character chain pattern. At this time, as shown in FIG. 15C, a character chain (A, A) connected to the delimiter code and a character chain (A, A) not connected to the delimiter code are created.

Step 5: When the search type is a forward match search, Step 6: From the search file, a character chain with a forward delimiter code corresponding to a character sequence with a forward delimiter code (a, a) The information is preferentially read, and then the character chain information with the backward delimiter code corresponding to the character chain with the backward delimiter code is read.

Step 7: If the search type is a backward match search, Step 8: Priority is given to the character chain information with the backward delimiter code corresponding to the character chain with the backward delimiter code (a, ashi) from the search file. Then, the character chain information with the forward delimiter code corresponding to the character chain with the forward delimiter code (a, a) is read. Step 9: Next, the character chain information without the delimiter code (a, a) is applicable. The character chain information without a delimiter code is read, and Step 10 to Step 16: collation is performed in the same procedure as in the first embodiment.

When reading the character chain information corresponding to the character chain with the forward delimiter code (A, and) from this search file, the area of the search file storage unit 11 where the character chain information with the forward delimiter code is stored. From (a, and) in the group containing the character type "a" (1,
1, 3, 1) can be easily obtained, and even when reading the character chain information of the character chain (a, shi) with a backward delimiter code, the character group (a) in the group including the character type "a" is read. Ah,
The character chain information corresponding to (c) can be easily obtained.

As described above, in the information retrieval apparatus of the third embodiment, a character chain pattern is created from the data to which the delimiter code is added without including the delimiter code, and the character chain information of the character chain pattern connected to the delimiter code is obtained. Depending on the connection position of the delimiter code, it is handled as character chain information with forward or backward delimiter code. The character chain information is stored in the search file for each group, for example, by classifying the character chain pattern into first characters. Therefore, the total number of character chain information in the search file is reduced, and the number of times of collation at the time of search can be reduced. In addition, since the information is arranged in groups, it is easy to read such information from the search file. In addition, since matching of the corresponding character chain information can be prioritized in accordance with a search type such as a forward match search, a backward match search, and a perfect match search, various searches can be performed at high speed.

(Fourth Embodiment) In the fourth embodiment, a search is performed in consideration of the hierarchy or order structure of data. This operation is performed in the information retrieval apparatus of FIG. 1 in addition to the components used in the first embodiment, a hierarchy / order-based delimiter adding unit 5 for adding a delimiter code according to a hierarchical structure to the end of each data, This is performed using the keyword-level-specific delimiter adding unit 15 that adds a delimiter code indicating the level to the keyword thus set.

In this case, the search file is created as shown in FIG.
Is performed according to the processing flow of

Steps 1 and 2: When the item (k-th) to be processed of the record data is designated, the record reading section 1 reads the record data from the record data storage section 2 and sets the record number.

Step 3: Hierarchical / order-based delimiter adding unit 5
Adds a hierarchical / sequential delimiter code to the beginning, between, and end of each data.

FIG. 18A shows an example of a plurality of data recorded in one item. Each data is assigned a level, and a delimiter code according to the level is added by the layer / order-based delimiter adding unit 5. FIG.
In (2), delimiter codes corresponding to levels 1, 2, and 3 correspond to $ S, $ T, and $ U.

The procedure of steps 4 to 16 is substantially the same as that of the first embodiment (FIG. 2, steps 4 to 15).
The character chain pattern generation unit 7 calculates the number of appearances for each character type in a character string to which a delimiter code for each layer and order is added, and creates a character chain pattern from this character string. At this time, the character chain pattern consisting only of the delimiter code is deleted (step 8). The search file generation unit 10 creates character chain information corresponding to a character chain pattern including various delimiter codes, stores it in the search file storage unit 11, and creates a search file. The search file storage unit 11 has a data structure shown in FIG.
Here, as shown in FIG. 18C, character chain information is stored. It should be noted that the character chain information is (first character occurrence times)
Number, second character appearance count, de - data number, is composed of a record number), character appearance number is sometimes hierarchy delimiter code for numbers. Here, delimiter codes are assigned levels according to the hierarchy, and character chain information is classified at this level. FIG. 18 (3) is an example of a search file for FIG. 18 (1).

Next, the flow of the retrieval process will be described with reference to FIG. FIG. 20 shows the flow of data at the time of retrieval.

Steps 1 to 3: As in the first embodiment, the keyword reading section 12 reads an input keyword, sets specified items, and specifies a search type. Here, as shown in FIG. 20A, it is assumed that “? Shibuya” is input as an input keyword and “2” is specified as its hierarchical level.

Step 4: The keyword layer-specific delimiter 15 selects a delimiter code for the specified layer. Step 5: The keyword delimiter 13 adds the selected delimiter code to the keyword according to the search type. I do. At this time, the keyword added with the delimiter is “? Shibuya @ T” (FIG. 20 (2)).

Step 6: Keyword character chain generation unit 17
Creates a character chain pattern by dividing the delimited keyword into character patterns (FIG. 20).
(3)).

The operations of the following steps 7 to 18 are as follows:
This is a process of acquiring character chain information corresponding to the divided character chain pattern, and is similar to the process of the first embodiment or the third embodiment.

FIG. 21 shows an example in which an order structure is adopted in data instead of hierarchizing the data. FIG. 21A shows an example of the data.
(2) shows the result of creating a search file. The procedure of search file creation and search processing is the same as in the case of hierarchization.

As described above, in the information retrieval apparatus of the fourth embodiment, when a plurality of data included in an item has a hierarchical or sequential structure, a delimiter code according to the hierarchical or sequential structure is added to the data. Create a search file,
At the time of search, the search is executed by specifying a hierarchy or an order. By doing so, it is possible to preferentially collate the information of the specified hierarchy or order from the character chain information stored in the search file, and to be aware of the hierarchical relation and the order relation in a plurality of data in one item. You can search.

[0095]

As is apparent from the above description of the embodiment, the information retrieval apparatus of the present invention can repeatedly register a plurality of data for one item which is a constituent element of the record data. Data such as a character string and a pattern including the specified keyword can be repeatedly searched at high speed within the item.

Even when a plurality of data have a common attribute, a plurality of data can be added in one item without newly increasing the number of items, so that registration is easy. Further, at the time of search, only one item needs to be specified, so that the search operation can be simplified and the search processing time can be shortened.

Further, since all the data in one item are registered in a distinguished manner, no search omission occurs.

Further, since a search within an item can be realized irrespective of the number of data added to one item, a high-speed search can be performed.

Also, since character chain information (search file) using a delimiter code is searched for by collating characters including a keyword, various searches such as a forward match search, a backward match search, a complete match search, and a partial match search are performed at high speed. Can be.

Further, since the character chain information including the delimiter code can be preferentially searched, it is possible to perform high-speed search processing, such as a perfect match search, which has conventionally taken a long time.

Further, since a plurality of data in one item can have a hierarchical relation or an order relation as a delimiter code, it is possible to register data having a hierarchical relation or an order relation in one record or one item. It is possible, and it is possible to perform a search in consideration of the hierarchical structure and the order relation.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an information search device according to an embodiment of the present invention;

FIG. 2 is a processing flow for generating a search file in the information search apparatus according to the first embodiment;

FIG. 3 is a data structure of a record data storage unit in the information search device according to the first embodiment;

FIG. 4 is a diagram showing character chain information of item j in the information search device of the first embodiment;

FIG. 5 is a data structure of a search file storage unit in the information search device according to the first embodiment;

FIG. 6 is a data flow until a search file is created in the information search device of the first embodiment;

FIG. 7 is a search processing flow in the information search device of the first embodiment,

FIG. 8 shows a search data flow in the information search device of the first embodiment;

FIG. 9 is a search processing flow in the information search apparatus of the second embodiment,

FIG. 10 shows a search data flow in the information search device of the second embodiment;

FIG. 11 is a processing flow for generating a search file in the information search apparatus according to the third embodiment;

FIG. 12 illustrates a data structure of a search file storage unit in the information search device according to the third embodiment;

FIG. 13 shows a data flow up to creation of a search file in the information search device of the third embodiment;

FIG. 14 is a flowchart of a search process in the information search device according to the third embodiment;

FIG. 15 shows a data flow at the time of search in the information search device of the third embodiment;

FIG. 16 is a processing flow for generating a search file in the information search apparatus according to the fourth embodiment;

FIG. 17 shows a data structure of a search file storage unit in the information search device of the fourth embodiment;

FIG. 18 is a data flow until a search file is created in the information search device of the fourth embodiment;

FIG. 19 is a search processing flow in an information search device according to a fourth embodiment;

FIG. 20 is a data flow at the time of search in the information search device of the fourth embodiment;

FIG. 21 is a data flow up to creation of a search file in the information search device of the fourth embodiment;

FIG. 22 shows the structure of record data in a conventional information search device;

FIG. 23 is a flowchart showing a data search procedure in a conventional information search device.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 record reading section 2 record data storage section 3 delimiter addition section 4 group delimiter addition section 5 layer / order delimiter addition section 6 priority indication section 7 character chain pattern generation section 8 delimitation determination section 9 number addition section 10 search file generation section 11 Search file storage unit 12 Keyword reading unit 13 Keyword delimiter addition unit 14 Keyword group delimiter addition unit 15 Keyword hierarchy / order delimiter addition unit 16 Priority collation unit 17 Keyword character chain generation unit 18 Collation unit 19 Search result display unit

──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Tadakazu Kikuchi 1006 Kazuma Kadoma, Osaka Prefecture Matsushita Electric Industrial Co., Ltd. (56) References JP-A-51-57259 (JP, A) 52226 (JP, A) JP-A-6-301721 (JP, A) JP-A-64-35627 (JP, A) JP-A-6-215041 (JP, A) Yoshifumi Masunaga, "Introduction to Relational Database-Data Model / SQL / Management System- ", Science Co., Ltd., March 10, 1993, first edition, 5th printing, p. 28-30 (58) Fields surveyed (Int. Cl. ⁷ , DB name) G06F 17/30 JICST file (JOIS)

Claims (8)

(57) [Claims] 1. An information retrieval apparatus for retrieving data including an input keyword from data in an item of record data having a plurality of items, wherein when the data in the item comprises a plurality of data, Delimiter assigning means for assigning a delimiter code between the beginning and end of the column connecting the data and the data; character chain pattern generating means for creating a character chain consisting of a set of two adjacent characters in the column; out in the data of each character constituting the
Search file creating means for creating character chain information including a current number , a record number, and the data number, storing the character chain information in a memory in association with the character chain, and creating a search file; Keyword-separating means for giving a character string; keyword-character-chain generating means for creating a character chain of keywords from keywords to which a delimiter code has been added; reading the character-chain information corresponding to the character chain of the keyword from the search file; An information retrieval apparatus, comprising: a collation unit configured to collate them and retrieve the data including a keyword. 2. The information according to claim 1, wherein said matching means is provided with priority matching means for prioritizing matching of character chain information corresponding to a character chain of a keyword including said delimiter code. Search device. 3. The information retrieval apparatus according to claim 1, wherein said character chain pattern generating means includes said delimiter code in said character chain. 4. The search file creation means according to claim 1, wherein the number of occurrences of the delimiter code in the character chain information is
4. The information search device according to claim 3, wherein a specific number is set. 5. The character chain pattern generating means creates the character chain without including the delimiter code, and the search file creating means determines a character string of a character chain adjacent to the delimiter code in the character chain. 2. The information retrieval apparatus according to claim 1, wherein the chain information is created separately from other character chain information. 6. The information search apparatus according to claim 5, wherein the search file creating means stores character chain information of a character chain adjacent to the delimiter code in the memory for each group. 7. The method according to claim 1, wherein the delimiter assigning unit assigns a plurality of types of delimiter codes indicating a hierarchy to the column, and the keyword delimiter assigning unit assigns a delimiter code indicating the hierarchy to an input keyword. Claim 1
An information retrieval device according to item 1. 8. The method according to claim 1, wherein the delimiter assigning unit assigns a plurality of types of delimiter codes indicating the order to the column, and the keyword delimiter assigning unit assigns the input keyword with the delimiter code indicating the order. Claim 1
An information retrieval device according to item 1.