US2007239708A1PendingUtilityA1

Database Search Method Of Searching A Document Database For A Desired Document File

Assignee: IBMPriority: Jan 10, 2003Filed: Jun 10, 2007Published: Oct 11, 2007
Est. expiryJan 10, 2023(expired)· nominal 20-yr term from priority
Y10S707/99931Y10S707/99933G06F 16/316
36
PatentIndex Score
0
Cited by
0
References
0
Claims

Abstract

An index file for use in search processing employing a search engine, retains information that indicates a correspondence of a keyword to its positional information. The index file comprises a key file that includes a list of pointers to character strings. The character strings are included in document files that are stored in a document database and to positional information, by document area where a character string in a document file appears and a POS file. The POS file includes a list of positional information containing information that specifies a document file where a character string exists and information that specifies a position of the character string in a document file regarding the character strings in the key file.

Claims

exact text as granted — not AI-modified
1 . A processor-implemented database search method, comprising: 
 storing a document file in a database;    controlling a transfer of a document file to and from the database;    searching the database on the basis of a keyword comprised of a character string and returning a search result;    using an index file for retaining information indicating a correspondence of the keyword to positional information of the keyword, based on a field in the document file that includes the keyword;    wherein the index file comprises a key file that includes a list of pointers to character strings and positional information;    wherein the character strings are included in the document file and are stored in the database by document area where a character string in the document file appears and by a positional information file;    wherein the positional information file includes a list of positional information containing information that specifies the document file where a character string exists and information that specifies a position of the character string in the document file regarding the character string in the key file; and    wherein for a variable-length chain included in the index file, a relational file is created by dividing each word in the variable-length chain into a plurality of fixed-length keywords.    
   
   
       2 . The database search method according to  claim 1 , further comprising: 
 reading the document file from the database, and extracting a text of the document file and information indicating a structure of the document file, for sending the text and the information to a search engine; and    the search engine creating the index file based on the text and the information indicating the structure of the document file.    
   
   
       3 . The database search method according to  claim 1 , wherein the index file retains, as the positional information on the keyword, information that identifies the document file containing the keyword and information that specifies a position of the keyword in the document file, each item of the positional information being associated with a corresponding item of the keyword listed as a separate item for a respective field.  
   
   
       4 . The database search method according to  claim 1 ,t wherein the index file comprises: 
 a first table that lists character strings contained in the document file stored in the database and pointers to positional information on the character strings, each character string being listed as a separate item for respective fields where the character string appears in the document file; and    a second table that lists positional information on each character string including the character strings listed in the first table, the positional information including information that identifies the document file containing the character string and information that specifies a position of the character strings in the document file.

Join the waitlist — get patent alerts

Track US2007239708A1 — get alerts on status changes and closely related new filings.

We store only your email — no account needed. See our privacy policy.