JPH11110402A - Document retrieving device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To speed up the retrieval by reducing the generation ratio of false drop at the time of retrieving document data. SOLUTION: At the time of retrieving a character string under a condition that forward coincidence and backward coincidence is obtained by using a processing part 2 of the document retrieving device, ' ' (a character indicating the head of a character string) and '$'(a character indicating the tail of a character string) are added as virtual characters for discriminating the head from the tail to retrieval character strings p1(='high speed') and p2 (='printer'), and characters to be retrieved t1 (='high speed printer'), t2 (='metropolitan speedway'), and t3 (='printer output'). Therefore, those character strings can be converted into p1 (= ' high speed'), p2' (='printer $'), t1' (=' high speed printer $'), t2' (='metropolitan speedway$'), t3' (=' printer output $'). Thus, signature can be extracted, and the generation ratio of false drop can be sharply reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document search apparatus, and more particularly, to a document search apparatus that searches for a document including a specified character string using a signature file.

[0002]

2. Description of the Related Art Conventionally, word processors and personal computers using word processing software have been used to create and edit documents by using a plurality of character types such as English and Japanese. A document search device is used to search for a partial character string in document data.

As this type of document search device, for example,
There is a document search device described in Japanese Patent Application Laid-Open No. 7-244671, and a document search is performed by using a binary bit pattern signature extracted from a certain character string by a certain method. 2 of this signature
The position of the bit at which "1" is set in the hexadecimal bit pattern is obtained by digitizing the characters and words constituting the character string and hashing the value from 0 to the maximum value of the bit position. . For example, if there is "copy" as a character string, "ko", "pi", "-"
Can be converted into numerical values "5", "7", and "12" using character codes, the numerical values indicate bit positions, and the bits are "5", "7", and "12" at the 12th position. 1 ", and becomes" 00000100001 ". The pattern of 0s and 1s is called a bitmap (bit pattern), and what is constituted by this bitmap is a signature.

[0004] The signature extraction method is described in "Ac
cess Method of Text '' (ChristosFaloutsos, Computing
Surveys, Vol. 17, No. 1, March 1985, pp. 49-74). According to this document, a signature called a word signature is created for each word constituting document data, and a superimposed signature is used as the signature of the document data. Here, the superimposition is an operation of calculating a logical sum of bit values at the same position in a plurality of signatures, and extracting a sequence of values of each logical sum as a new signature.

Further, in order to search for a partial character string constituting a word, the document data is divided into a character string having a fixed number of characters while having an overlapping portion, and each of the document data is divided into character strings in the same manner as the above-mentioned word signature. There is a way to superimpose a string signature.

There is also a method in which longer document data is divided into logical blocks such as sentences and paragraphs, and a plurality of signatures extracted from each block correspond to one document. Here, the signature extracted from the block is called a block signature.

When a signature is used to search for a document including a search character string, the same bit pattern signature may be extracted from different character strings. May be detected. This document is called False Drop. On the other hand, a document including a search character string is called an actual drop.

In a conventional document retrieval apparatus, signatures are extracted for each document, and each signature is stored collectively in a file called a signature file.
Signature files are roughly classified into two types according to the signature storage method. One method is to simply store signatures in order. The file configuration by this method is called a sequential configuration. The other is
In this method, each bit of the signature is stored in a separate bitmap for each bit position. The file configuration by this method is called a bit slice configuration. The signature file with the bit slice configuration is “Partial-Ma
tch Retrieval via Method of Superimposed Codes ''
(Charles S. Roberts, Proceedings of the IEEE. Vo
l, 67, No. 12, December, 1979, pp 1624-1979).

A method for compressing a bitmap of a signature having a sequential structure is described in “Description and
Performance Analysis of Signature File Methods for
Office Filing "(Christos Faloutsos, ACM Transacti
on Office Information Systems, Vol.5, No.3, July 198
7, pp. 237 to 257), there is a method using run-length coating or the like.

(1) Therefore, when a character string is searched by using the above-described conventional document search apparatus on the condition of a head match and a tail match, the search character strings with the head match and the tail match are p1 and p2. And the character strings to be searched are t1, t
This will be described in the following example in which 2, t3 is set. p1 = “high-speed” p2 = “printer” t1 = “high-speed printer” t2 = “capital expressway” t3 = “printer output” In the case of using the signature extracted from the above search character strings p1 and p2, generated characters Cannot be distinguished from the beginning or end of the character string.
When a search was performed on the signatures extracted from 1, t2, and t3, t1 and t2 were the search results under the condition of p1, and t1 and t3 were the search results under the condition of p2.

(2) Further, in the conventional document retrieval apparatus, the document data is divided into one or more partial character strings, and a signature corresponding to each partial character string is extracted. This partial character string was called a block, and the number of characters in each block was constant. Therefore, the number of characters in the last block of each document data is equal to or less than the number of characters in the blocks other than the last block.

(3) Further, in the conventional document search device,
As a way to extract the signature, the adjacent N
An N-gram method has been used in which the position of a bit whose value should be set to 1 by signature is determined according to a character string of a character (N is an integer of 2 or more). For example, in the 3-gram system, a character string "Ricoh's high-performance camera" is decomposed into the following partial character strings to extract a signature. Ricoh / Co / High / High performance / High performance / Performance / Noh camera / Camera

(4) In the conventional document search device,
When searching by a conditional expression that includes multiple search strings connected by the logical operators AND and OR, such as “high-speed or printer”, scan the signature file for each search string and search for each. By performing a set operation on the results, a final search result has been obtained.

[0014]

However, in such a conventional document search apparatus, (1) the search character strings described above are p1 and p2, and the character strings to be searched are t1 and p1.
In the case of t2 and t3, only the identity of the character string can be determined, and it is not possible to distinguish whether the character string matches at the beginning or at the end of the character string even under the condition of head match and tail match. Therefore, the occurrence rate of false drop, which is a document that does not include a search character string as a search result, becomes high, and there is a disadvantage that it takes time to search.

In the conventional document retrieval apparatus of (2), if the document data is divided into partial character strings with a fixed number of characters and the signature is extracted, the number of characters in the block at the end of the document data becomes uneven. However, there is an inconvenience that false drops increase and search takes time.

Further, in the conventional document search apparatus of the above (3), if a partial character string is simply cut out from adjacent N characters, it is designated as a search character string such as "-high" in the above example. Since the signature bits are assigned to a character string that is rarely performed, the number of bits that are 1 in the signature increases, so that the probability that a false drop occurs and the search takes time is disadvantageous. there were.

In the conventional document search apparatus of the above (4), when a search is performed using a conditional expression including a plurality of search character strings, the number of search character strings included in the conditional expression increases. As a result, the number of scans increases, so that the search time becomes longer and the search performance decreases.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned disadvantages of the related art, and an object of the invention described in claim 1 is to reduce a false drop occurrence rate in a document search on the basis of a head match and a tail match. It is an object of the present invention to provide a document search device that can reduce the speed and speed up the search.

An object of the invention described in claim 2 is to provide a document search apparatus capable of suppressing an increase in false drops that occur when extracting a block signature from document data, thereby achieving a high-speed search. It is in.

An object of the invention described in claim 3 is to extract false signatures from document data having many types of characters by the N-gram method by suppressing an increase in false drops and speeding up retrieval. A document search device is provided.

It is an object of the present invention to provide a document search apparatus capable of speeding up a search when searching for a document using a conditional expression including a plurality of search character strings.

[0022]

According to a first aspect of the present invention, there is provided an input unit for inputting document data and a search character string of a registered document, and a character code for inputting the document data and the search character string input to the input unit. A data unit having a processing unit that converts the data into a signature and extracts the signature, a signature file that stores the signature extracted by the processing unit, and a record file that obtains the document data by referring to the record corresponding to the record identifier; and In the document search device including an output unit that outputs data, the processing unit extracts virtual characters representing the beginning or end of each character string before extracting the signature of the document data and the search character string. Extract the signature by adding it, and distinguish whether the same substring appears at the beginning or end of the character string with the signature. The features.

According to this, before the signature of the document data and the search character string is extracted in the processing unit, virtual characters representing the head or end of each character string are added to the document data or the search character string. Then, the signature is extracted, so that it is possible to determine whether or not the substring to be searched appears at the beginning or end of the character string using only the signature. , Search performance and search speed can be improved.

According to a second aspect of the present invention, there is provided an input unit for inputting document data and a search character string of a registered document, and converting the document data and the search character string input to the input unit into character codes to generate a signature. A data unit having a processing unit for extracting, a signature file for storing the signature extracted by the processing unit, and a record file for obtaining document data by referring to a record corresponding to the record identifier; and an output unit for outputting document data Wherein the processing unit divides the document data into a plurality of blocks and, when extracting a signature corresponding to each block, divides the document data so that the number of characters in each block is uniform. It is characterized by the following.

According to this, the processing unit divides the document data into a plurality of blocks and, when extracting a block signature corresponding to each block, divides the block so that the number of characters in each block is uniform. As a result, it is less likely that more bits than necessary are allocated or the number of bits is insufficient, so that an increase in false drops is suppressed and search can be speeded up.

According to a third aspect of the present invention, there is provided an input unit for inputting document data and a search character string of a registered document, and converting the document data and the search character string input to the input unit into a character code to generate a signature. A data unit having a processing unit for extracting, a signature file for storing the signature extracted by the processing unit, and a record file for obtaining document data by referring to a record corresponding to the record identifier; and an output unit for outputting document data The processing unit, when extracting a signature according to a character string of adjacent N characters (N is an integer of 2 or more) of the document data, Character strings are of the same character type.

According to this, when the signature is extracted in the processing unit in accordance with the character string of the adjacent N characters of the document data, the character strings of the adjacent N characters are of the same character type. Since the number of types of character strings can be reduced as compared with the conventional method, an increase in false drops is suppressed, and search can be speeded up.

According to a fourth aspect of the present invention, there is provided an input unit for inputting document data and a search character string of a registered document, and converting the document data and the search character string input to the input unit into character codes to generate a signature. A data unit having a processing unit for extracting, a signature file for storing the signature extracted by the processing unit, and a record file for obtaining document data by referring to a record corresponding to the record identifier; and an output unit for outputting document data The processing unit, when searching for a character string based on a conditional expression including a plurality of search character strings, extracts the signature of all search character strings included in the conditional expression Then, a cursor for scanning a bit map bit-sliced in accordance with the bit whose signature value is 1 is prepared. And performing a search while scanning in parallel in accordance with the contents of the conditional expression.

According to this, when the processing unit searches for a character string based on a conditional expression including a plurality of search character strings, signatures of all search character strings included in the conditional expression are extracted and To prepare a cursor that scans a bitmap that is bit-sliced according to the bit whose signature value is 1 and perform a search while scanning the cursor in parallel according to the contents of the conditional expression, search characters included in the conditional expression Even if the number of columns increases, search results can be obtained only by scanning the signature file once, and the search can be speeded up.

[0030]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments 1 to 4 of the present invention will be described below with reference to the drawings.

Before describing each embodiment, a schematic configuration of a document retrieval apparatus common to each embodiment will be described. FIG. 1 is a block diagram showing a schematic configuration of the document search device according to the present embodiment. In FIG. 1, the document search device includes an input unit 1, a processing unit 2, a character string input processing unit 3, a record identifier calculation processing unit 4, a document search processing unit 5,
It comprises a document output processing unit 6, a storage position calculation processing unit 7, a document registration processing unit 8, a signature extraction processing unit 9, an output unit 10, a data unit 11, a signature file 12, a record file 13, and the like.

The target search character string and the character code of the document data of the registered document are ASCII (American Standalone).
ard Code Information Interchange), even if the number of bytes of each character is 1 byte, EUC
1 byte as in (Extended Unix Interchange), 2 bytes
It may be a character code in which a byte and a 3-byte character are mixed. The character code for internal processing used in the character string input processing unit is a character code to be converted (for example, EU).
It need not be the same as C). Also, in the present invention, a case where EUC is used will be described.

As shown in FIG. 1, the search character string input from the input unit 1 and the document data of the registered document are converted from the character code for input by the character string input processing unit 3 of the processing unit 2 to E.
Converted to UC. In addition, document data that is an actual drop at the time of retrieval is converted from the EUC to a character code for output by the document output processing unit 6. Therefore, the search character string and the document data of the registered document are
It is always processed as an EUC character string, and the document data is always stored in the record file 13 of the data section 11 as an EUC character string.

As described above, the document search apparatus according to the present embodiment has a function of registering and searching for a document, and is composed of an input unit 1, a processing unit 2, a data unit 11, and an output unit 10. Have been. The processing unit 2 includes a character string input processing unit 3, a signature extraction processing unit 9, a document search processing unit 5, a record identifier calculation processing unit 4, and a document output processing unit 6.
And a document registration processing unit 8 and a storage position calculation processing unit 7, and further include a cursor movement processing unit, a conditional expression clearing unit, and a conditional expression prospect calculation unit (not shown).

The data section 11 includes a signature file 12 and a record file 13.
When registering a document, the input unit 1 accepts input of document data to be registered. The input document data is converted into a predetermined character code by a character string input processing unit 3 and then transferred to a document registration processing unit 8. . The document registration processing unit 8 stores the first
Then, the document data is stored in the record file 13 of the data section 11. Second, the storage location of the signature is calculated from the identifier of the record storing the document data by using the storage location calculation processing unit 7. Third, the document data is divided into blocks each having a fixed number of characters, and a block signature is extracted from each block using the signature extraction processing unit 9. However, adjacent blocks have overlapping portions with a certain number of characters. Fourth, the block signature is stored at a predetermined storage location in the signature file 12 of the data section 11.

Regarding the document search operation, first,
A search character string is received by the input unit 1, and the input search character string is converted into a predetermined character code by the character string input processing unit 3, and is then passed to the document search processing unit 5. First, the document search processing unit 5 extracts a partial character string within the limited number of characters from the search character string, extracts the signature from the partial character string using the signature extraction processing unit 9, and extracts the signature. Is the signature of the search string. Here, the limited number of characters refers to the number of characters in an overlapping portion between blocks constituting registered document data.
Secondly, the bit for which "1" is set in the search signature is checked, and in the signature file of the data section 11, the bit map corresponding to the previously checked bit is referenced, and the search character string is searched. The storage position of the block signature extracted from the block that is determined to contain the is calculated.

Third, the value of the record identifier is obtained from the storage location of the block signature using the record identifier calculation processing unit 4. Fourth, document data is obtained by referring to the record corresponding to the record identifier in the record file 12 of the data section 11. Fifth, the obtained document data is passed to the document output processing unit 6. The document output processing unit 6 checks whether or not the passed document data really includes the search character string, removes the false drop, converts the actual drop into a predetermined character code,
The document data is passed to the output unit 10. The output unit 10 outputs the passed document data.

(Embodiment 1) In the first embodiment, when a character string is searched by using the processing unit 2 of the document search apparatus on the basis of a head match and a tail match, a head match and a tail match are required. The following search character strings p1 and p2 and character strings t1, t2 and t3 to be searched are set. p1 = “high speed” p2 = “printer” t1 = “high speed printer” t2 = “capital expressway” t3 = “printer output”

In these character strings, virtual characters for discriminating between the beginning and the end are represented by "@" (a character indicating the beginning of the character string) and "@" (a character indicating the end of the character string). Character), and convert it to the following character string before extracting the signature. p1 '= "＾ high speed"p2' = "printer" t1 '= "＾ high speed printer"t2' = "＾ capital expressway" t3 '= "＾ printer output"

As described above, in the first embodiment, the signature is extracted including the virtual character for distinguishing the head or the end of the character string, so that the search character strings p1 ′ and p2 In any case, the search result based on 'can search only for "high-speed printer" at t1'. This means that, as in the conventional example, the search results are t1 and t2 under the condition of p1, and t1 and t2 under the condition of p2.
Compared to the case where 1, t3 is a search result, the occurrence rate of false drops can be greatly reduced, and as a result, search performance and search speed can be improved.

(Embodiment 2) In Embodiment 2, the document data is divided into one or more partial character strings (blocks) using the processing unit 2 of the document search apparatus, and a signature corresponding to each block is obtained. When extracting, each block is divided so that the number of characters in each block is constant. Referring to FIG. 2, FIG. 2A shows a comparative example in which document data is divided by a fixed number of characters, and FIG. 2B shows that the number of characters in each block is equal to or less than the appropriate number of characters. Embodiment 2 in which each block is divided such that the number of characters in each block is uniform.
This is an example. As shown in FIG. 2A, when the number of characters of a block to be divided is determined in advance and the number of characters is divided, the number of characters of the last block (block 4 in the figure) for each document data is changed to the other blocks (blocks in the figure) 1,
2,3) or less. In such a situation, with respect to the number of bits of the signature, let L be the number of characters in an appropriate block, L1 be the number of characters in blocks 1, 2, and 3,
Assuming that the number of characters in block 4 is L2, their magnitude relationship may be the following inequality relationship. L2 ≦ L ≦ L1

In the case of the above inequality relationship, in the block signature, more bits than necessary are allocated to the block 4 while the number of bits is insufficient for the blocks 1, 2, and 3. Therefore, the false drop was increased.

On the other hand, in Embodiment 2, FIG.
As shown in (b), the number of characters in each block is equal to or less than L, the number of characters in each block is equalized, the document data is divided, and the signature is extracted. Is possible, and a false drop can be reduced.

(Embodiment 3) In this embodiment 3, the processing unit 2 of the document search apparatus is used to output kanji characters such as Japanese characters.
When extracting a signature from document data having many character types such as Hiragana and Katakana using the above-described N-gram method, the character type is not cut out by arbitrarily adjacent N characters (N is an integer of 2 or more). The same arbitrarily adjacent N characters are cut out to extract a signature. By doing so, the character string “Ricoh's high-performance camera” in the example shown above can be cut out as “Ricoh / high-performance / camera”. Compared to / high / high performance / high performance / high power / high power / camera, extra cutting such as (high / high power / high power) and (high power / high power) On the other hand, there is no need to assign a signature bit, so that the probability of a false drop occurring can be greatly reduced.

As described above, according to the third embodiment, N
When extracting signatures by extracting document data using the -gram method, only the same character type is extracted, so that the types of extracted character strings can be significantly reduced as compared with the conventional case. As described above, if the types of character strings are sufficiently reduced, the length of the signature (the number of bits) can be made to match the type. With such a signature, if the search string is N characters, then the bitmap to be scanned only needs to be one bit slice, and no false drop occurs.

In the document search apparatus according to the third embodiment, if the search character string is "printer",
If the gram method is used, the signature will be submitted from a character string that is unlikely to be a search character string, as in pudding / linta / inter, but in such a case, only the above-mentioned 3-gram method is used. Instead of 4-gr
When used in combination with another method such as the am method or the 5-gram method, the same advantageous effects as in the above case can be obtained.

(Embodiment 4) In Embodiment 4, a plurality of search character strings connected by logical operators AND and OR, such as "high speed or printer", are used by using the processing unit 2 of the document search apparatus. Thus, even when a search is performed using a conditional expression that includes, the search can be speeded up.
That is, first, a signature is extracted from all the search character strings (patterns) included in the conditional expression, and secondly, a bit map obtained by bit-slicing according to the bit whose extracted signature value is 1 is obtained. Third, by preparing a plurality of cursors to be scanned and scanning them while moving them in parallel according to the contents of the conditional expression, the signature file 12 is scanned only once.
Search results can be obtained. Moreover, since the cursor always scans in the forward direction, the MH (Modified Huffm
There is an advantage that bitmaps compressed by the an) method can be handled.

FIG. 3 is a diagram showing the relationship between the above-described search character string pattern, the cursor, and the bit-sliced bitmap. Here, the cursor has a one-to-one correspondence with the bit-sliced bitmap.
In other words, among the signatures extracted from the search character string included in the conditional expression, only one cursor is prepared even if there is a plurality of signatures having a value of 1 at a certain bit position. By doing so, even if the number of search character strings included in the conditional expression increases, scanning of the bit-sliced bitmap can be performed by one.
It can be done in times.

The cursor is associated with each search character string (pattern). In addition, the cursors are arranged in order of search efficiency for each pattern. Here, the order in which the search efficiency is high is an order in which the number of bits having a value of 1 is small in a bitmap bit-sliced according to the cursor. Thereby, search efficiency can be improved.

In each search character string, the first cursor is called the primary cursor. At the time of the search, the primary cursor having the smallest bitmap position is selected from the primary cursors. Search for a bit.

Then, after the movement of the primary cursor,
If it is assumed that a signature is found at the position indicated by each primary cursor, it is determined whether or not the document data from which those signatures are extracted satisfies the conditional expression.

The conditional expression is converted in advance into a conjunction standard form. The conjunction standard form represents a form of a logical expression in which a literal is a disjunction (sum) of a conjunction (product). Here, a literal is associated with one search character string. For example, the conditional expression ((high-speed OR high-definition) AND printer)
It is converted to the conjunction standard form of “(high-speed AND printer) OR (high-definition AND printer)”. in this way,
By setting the conditional expression to the conjunctive normal form, it is possible to determine whether the entire conditional expression is true or false based on whether any of the conjunctions is true or false.

When the primary cursor corresponding to a certain search character string reaches the end of the bit-sliced bitmap, the conjunction that uses the search character string corresponding to the primary cursor as a literal is removed.
Conditional expressions can be simplified. In other words, if it is found that the signature of the search string “high performance” cannot be found, then the conditional expression is changed to “high-speed AND printer”.
What should be searched for.

Assuming that signatures corresponding to the respective primary cursors are found at the positions indicated by the primary cursors, and when it is determined that the document data from which those signatures are extracted satisfies the conditional expression, if the signature data is determined to satisfy the conditional expression, It is checked whether the value of the bit position is “1” for the cursors after the primary cursor, and it is determined whether or not the conditional expression is satisfied.

As described above, the search result is obtained while repeating the cursor movement. The details are shown in FIGS. 4 to 7 of the flowchart.
These flowcharts are flowcharts of the cursor movement processing unit (FIGS. 4 and 5), the conditional expression settlement unit (FIG. 6), and the conditional expression calculation unit (FIG. 7). And the conditional expression expectation calculation unit are called. In each of these figures, the search character string that constitutes the conditional expression is referred to as a pattern, and each of those patterns has position, checke
d and match. The initial values of position and checked are “0”, and “0” represents a virtual head position of the bit sliced bitmap. Also matc
The initial value of h is FALSE.

Regarding the function search in step 45 in FIG. 4, the cursor designated by the argument is advanced to the bit whose next value is "1", and the bit position is returned. However, if a bit whose next value is “1” is not found, a negative value is returned. Accordingly, it is determined whether to go to step 47 or step 53 depending on whether h is positive or negative in step 46.

In these flowcharts, the search result and the document data serving as candidates are represented by t. If t is undefined, its value is set to null (step 49). The attribute values of t include start and end, and represent the start and end of the storage position of the signature according to t.

In step 68 of FIG. 5, a set of signatures corresponding to t is expressed as a family signature.

As described above, when the processing unit 2 of the document search apparatus according to the fourth embodiment searches for a character string based on a conditional expression including a plurality of search character strings, all of the conditions included in the conditional expression Extract the signature of the search character string, prepare a cursor that scans the bitmap that is bit-sliced according to the bit whose signature value is 1, and search while scanning the cursor in parallel according to the contents of the conditional expression Therefore, even if the number of search character strings included in the conditional expression increases, the search result can be obtained only by scanning the signature file once, and the search can be speeded up.

Although the document search apparatus according to the present embodiment has been described for a case where a document search is performed, the present invention is not limited to this, and can be applied to a more general search apparatus. It is. For example, gender, blood type,
Allocate a bit to distinguish the prefecture of origin from the signature, and then select "sex = male AND blood type = AB AND (origin = Saitama
OR hometown = Yamanashi) ”, it is possible to search for the contents of this conditional expression at high speed.

[0061]

As described above, according to the first aspect of the present invention, the false drop occurrence rate is reduced in the document search on the condition of the head match and the tail match, thereby speeding up the search. be able to.

According to the second aspect of the present invention, it is possible to suppress an increase in false drop that occurs when extracting a block signature from document data, and to speed up a search.

According to the third aspect of the present invention, when extracting signatures from document data having many types of characters by the N-gram method, an increase in false drops can be suppressed, and the search can be speeded up. .

According to the fourth aspect of the present invention, it is possible to speed up a search when searching for a document using a conditional expression including a plurality of search character strings.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of a document search device according to the present embodiment.

FIG. 2 is a diagram illustrating a state of division of document data when extracting block signatures by dividing the document data according to a second embodiment.

FIG. 3 is a diagram showing a relationship between a search character string pattern, a cursor, and a bit-sliced bitmap according to a fourth embodiment.

FIG. 4 is a flowchart of a cursor movement processing unit according to the fourth embodiment.

FIG. 5 is a flowchart of a cursor movement processing unit according to the fourth embodiment.

FIG. 6 is a flowchart of a conditional formula settlement unit according to a fourth embodiment.

FIG. 7 is a flowchart of a conditional expression prospect calculation unit according to the fourth embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Input part 2 Processing part 10 Output part 11 Data part

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Hideaki Nakayama 1-3-6 Nakamagome, Ota-ku, Tokyo Inside Ricoh Co., Ltd.

Claims (4)

[Claims] An input unit for inputting document data and a search character string of a registered document; a processing unit for converting the document data and the search character string input to the input unit into a character code to extract a signature; A document search device comprising: a data unit having a signature file storing a signature extracted by a processing unit and a record file for obtaining document data by referring to a record corresponding to a record identifier; and an output unit outputting the document data In the processing unit, before extracting the signature of the document data and the search character string, extract the signature by adding virtual characters representing the beginning or end of each character string, the same partial character string A document search device characterized in that whether or not it appears at the beginning or end of a character string is distinguished by a signature. 2. An input unit for inputting document data and a search character string of a registered document, a processing unit for converting the document data and the search character string input to the input unit into a character code and extracting a signature, A document search device comprising: a data unit having a signature file storing a signature extracted by a processing unit and a record file for obtaining document data by referring to a record corresponding to a record identifier; and an output unit outputting the document data In the document search, the processing unit may divide the document data into a plurality of blocks, and extract a signature corresponding to each block so that the number of characters in each block is uniform. apparatus. 3. An input unit for inputting document data and a search character string of a registered document, a processing unit for converting the document data and the search character string input to the input unit into a character code and extracting a signature, A document search device comprising: a data unit having a signature file storing a signature extracted by a processing unit and a record file for obtaining document data by referring to a record corresponding to a record identifier; and an output unit outputting the document data In the processing unit, when extracting a signature according to a character string of adjacent N characters (N is an integer of 2 or more) of the document data, each character string of the adjacent N characters may have the same character type. A document search device characterized by the following. An input unit for inputting document data and a search character string of the registered document; a processing unit for converting the document data and the search character string input to the input unit into a character code to extract a signature; A document search device comprising: a data unit having a signature file storing a signature extracted by a processing unit and a record file for obtaining document data by referring to a record corresponding to a record identifier; and an output unit outputting the document data In the processing unit, when searching for a character string based on a conditional expression including a plurality of search character strings, extract the signatures of all search character strings included in the conditional expression, the value of each signature A cursor for scanning a bit map bit-sliced according to the bit being 1 is prepared, and each of the cursors is moved according to the contents of the conditional expression. A document search device for performing a search while scanning in parallel.