Hitachi, Ltd.

Hitachi Maxell, Limited

Hitachi Computer Engineering Co., Ltd.

Hisamitsu Kawaguchi

Kanji Kato

Atsushi Hatakeyama

Hiromichi Fujisawa

Noriyuki Yamasaki

Toru Takahashi

Katsumi Tada

Mitsuru Akizawa

Masaaki Fujinawa

Satoshi Asakawa

Hidefumi Masuzaki

Masaharu Murakami

Noriyuki Kaneoka

Mitsuo Ooyama

Takuya Okamoto

Natsuko Mizutani

Hirofumi Yamashita

Kouki Noguchi

Takanori Shimura

Natsuko Sugaya

Masatsugu Shinozaki

Akira Kito

Hidenori Yamada

Hiroe Ando

Naoki Minegishi

A text cataloging method includes a step of cataloging already-analyzed-text data obtained from an analysis of a logical structure of a text to be cataloged in a text database, a step of creating a structure index by sequentially superposing logical structures of texts to be cataloged, wherein a single metaelement is used for representing a group of elements in the texts having the same position of appearance in one of the texts and the same element type, a single piece of meta-character-string data is used for representing a group of pieces of character-string data in the texts having the same position of appearance in one of the texts, and a context identifier is assigned to each metanode composing a tree-like structure of the structure index for uniquely identifying the metanode; a step of generating structured-full-text data composed of definitions of associative relations between all pieces of character-string data included in already-analyzed-text data of each text to be cataloged, and context identifiers of pieces of meta-character-string data in the structure index used for representing the pieces of character-string data; and a character-string-index updating step, including the sub-steps of extracting partial character strings, generating structured-character-position information, and updating a character-string index.

Structured-text cataloging method, structured-text searching method, and portable medium used in the methods

A document search method and apparatus and a portable medium used therefor are described, in which when registering a document in a data base, the logic structures of each document to be registered are superposed one on another to generate a structure index in which the structure elements having the same position of occurrence in the document are represented by a single meta-node. At the time of document search, a mass of the meta-nodes meeting a specified structural condition is determined with reference to the structure index. A string index is searched with the meta-node identifiers as a key thereby to determine a mass of documents meeting the specified condition. As a result, a highly accurate structure-specified search is made possible on a document data base including a mass of structured documents. In the structure-specified search of structured documents, the conditions for the position of occurrence of the logic elements in the document are specified, thereby making possible a highly accurate structure-specified search.

Document search method for registering documents, generating a structure index with elements having position of occurrence in documents represented by meta-nodes

Document search method and apparatus and portable medium used therefor

A text cataloging method includes a step of cataloging already-analyzed-text data obtained from an analysis of a logical structure of a text to be cataloged in a text database, a step of creating a structure index by sequentially superposing logical structures of texts to be cataloged, wherein a single metaelement is used for representing a group of elements in the texts having the same position of appearance in one of the texts and the same element type, a single piece of meta-character-string data is used for representing a group of pieces of character-string data in the texts having the same position of appearance in one of the texts, and a context identifier is assigned to each metanode composing a tree-like structure of the structure index for uniquely identifying the metanode; a step of generating structured-full-text data composed of definitions of associative relations between all pieces of character-string data-included in already-analyzed-text data of each text to be cataloged, and context identifiers of pieces of meta-character-string data in the structure index used for representing the pieces of character-string data; and a character-string-index updating step, including the sub-steps of extracting partial character strings, generating structured-character-position information, and updating a character-string index.

Structured-text cataloging method, structured-text searching method, and

A method for making document information searches. In performing a document search with respect to the desired key word, two stages of presearch are carried out. In a first stage of presearch, a character component table in which an existence of character codes for every document is stated with respect to all the character codes contained in the group of document text data of stored documents is generated, and the character component table is searched for all the character strings constituting a desiredly designated search subject key word to thereby extract all the documents each containing all the character codes constituting the search subject key word. In a second stage of presearch, contracted text data for every document in which adjuncts and duplication of repeatedly stated words contained in advance in the text data are eliminated is generated, and the documents each containing the search subject key words by word are extracted from the documents extracted by the first presearch. After the second stage of presearch, text search is performed in accordance with a neighbor condition, a contextual condition, or the like.

Presearch type document search method and apparatus

A method of fast clipping, despite of large number of users, can be achieved through analyzing query expressions, storing the number of query terms included in the query expressions in a term number count table, generating a finite automaton for matching the terms occurring in text data with all terms included in the query expressions, generating a user identifier table for storing the identifiers of users in association with the terms included in the query expressions, matching the terms by scanning the text data by the finite automaton, calculating for each user the occurrence count of terms occurring in the text data as substrings coincident with the terms included in the query expressions made to the user identifier table, storing the calculated occurrence count in the term occurrence count region of the table, comparing the calculated term occurrence count of the table with the number of terms included the query expressions, and when a match is found from the comparison, delivering the text data to the user.

News clipping method and system

A document retrieval method and system for retrieving, from a document database storing document data in the form of character codes, a document which contains given search terms and which meets a given search query condition. From documents loaded from the document database, a document containing terms which match the search terms is searched to generate document identification (ID) information including a document identifier of the searched document and containing match terms found to match with the search terms as well as term identifiers of the match terms and position information of the match terms in the searched document. A decision is then made as to whether or not the position information of the match terms satisfies a positional condition specified in the search query condition concerning a positional relation between the search terms, and match information is then generated indicating satisfaction of the search query condition when the positional condition is satisfied. Through a proximity condition decision, it is ascertained whether the match terms satisfy an inter-term distance condition specified in the search query condition. Through a contextual condition decision, it is determined whether the match terms satisfy a concurrence condition specifying concurrence of the search terms in a same sub-sentence, a same sentence or a same paragraph. Through a logical condition, it is decided whether the match terms satisfy a logical condition between the search terms specified in the search query condition.

Document retrieval method and system

A neighboring plural-character occurrence bitmap of a practical capacity capable of eliminating noises by hashing is realized, and a high speed full text search is realized equivalently, by greatly reducing the number of documents to be searched even if a search term constituted by a combination of English characters and words is used. Text data is segmented into words, and n-character strings at every (m+l)-th character positions are extracted from each word. A neighboring plural-character occurrence bitmap is created which stores data representing a presence of each neighboring plural-character string at a certain entry thereof. N-character strings at every (m+l)-th character positions are extracted from a search term and the neighboring plural-character occurrence bitmap is searched by using a search control program. Since the neighboring plural-character occurrence bitmap is searched prior to searching condensed texts, documents not relevant to the search term can be discarded and a high speed full text search can be realized.

Document search method wherein stored documents and search queries

A text search method for structured documents and apparatus wherein structured document database to be searched is created by adding logical structure length information to logical structure discriminating information on the basis of a structured document constituted by a plurality of logical structures and logical structure discriminator information, and in searching the structured document database in accordance with an entered search key constituted by a logical structure and a search character string, the search of a logical structure other than the entered logical structure is skipped based upon the data length information. In the method and apparatus, by using a structured document, structured document database to be searched is created which is constituted by a condensed text for each logical structure constituted by a list of words contained in the logical structure, a character occurrence bitmap for each logical structure constituted by a list of characters contained in the logical structure and an source structured document, and then the structured document database is searched in accordance with an entered search key constituted by a logical structure and a search character string. An source structured document is searched optionally depending upon the entered search key.

Text search method and apparatus for structured documents

A typical structure of a file server system is a file server system having a plurality of file servers connected in parallel on a network and sharing files placed distributedly in the file servers among a plurality of client computers, and there are provided in a specific file server among the plurality of file servers, a load information monitoring device for measuring respective loads of the plurality of file servers and a file access request distributing device for referring to the loads measured by the load information monitoring device so as to select a file server having a light load from the plurality of file servers having light loads, and distributing a file access request transmitted from client computers to the selected file server.

File server system and file access control method of the same

A method and apparatus for performing a document information search to uncover specified text data containing a given search subject key word from a group of document text data stored in a memory. In the document information search method, two stages of presearch are carried out to perform the document search with respect to a desired subject key word. In a first stage of presearch, a character component table is generated in which the existence of character codes for every document is set forth with respect to all the character codes contained in the group of document text data of stored documents. The character component table is searched for all the character codes comprising a designated search subject key word to thereby extract all documents containing all the character codes comprising the search subject key word. Further, in the presearch step, all texts without the possibility of containing the search subject key word are eliminated. A comprehensive, narrowed text search is thereby performed in accordance with the search subject key word.

Growing community of inventors

Sagamihara, Japan

Hisamitsu Kawaguchi