SEARCH PROCESSING METHOD AND SEARCH SYSTEM

• US 20080097993A1
• Filed: 04/13/2007
• Published: 04/24/2008
• Est. Priority Date: 10/19/2006
• Status: Active Grant

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Nanba, Isao
• Granted Patent: US 7,680,852 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3331 Query processing

Abstract

A technique to maintain the fast search capability for the large-scale documents without causing the update delay is provided. This search processing method includes: causing an index search unit for carrying out an index search using a search index before document update to carry out the index search relating to a search request, and obtaining a first list of document IDs of pertinent documents; causing a string pattern matching unit having document contents after the document update to carry out a string pattern matching relating to the search request for the document contents after the document update, and obtaining a second list of document IDs of pertinent documents; and generating a search result for the search request by using the first and second lists and a third list of document IDs of documents relating to the document update.

Claims

1. A search processing method, comprising:

causing an index search unit for carrying out an index search using a search index before a document update to carry out said index search relating to a search request, and obtaining a first list of document IDs of pertinent documents;

causing a string pattern matching unit having document contents after said document update to carry out a string pattern matching relating to said search request for said document contents after said document update, and obtaining a second list of document IDs of pertinent documents; and

generating a search result for said search request by using said first and second lists and a third list of document IDs of documents relating to said document update.

3. The search processing method as set forth in claim 1, wherein said causing said index search unit comprises carrying out a normalization processing for said index search against a search word relating to said search request, and

said causing said string pattern matching unit comprises carrying out a normalization processing for said string pattern matching against said search word relating to said search request.

2. The search processing method as set forth in claim 2, wherein said generating comprises:

generating a fourth list by removing said document IDs included in said third list from said first list; and

generating a sum-set of said fourth list and said second list.

4. A search system, comprising:

a first search index used to carry out an index search in response to a search request;

a first document ID manager, for added or modified document, that is used in search and stores document IDs of updated documents; and

a first pattern matching document storage that stores data of said updated documents to carry out a character string in said search, and

wherein said index search relating to said search request is carried out by using said first search index to obtain a first list of document IDs of pertinent documents,

a string pattern matching relating to said search request is carried out for said first pattern matching document storage to obtain a second list of document IDs of pertinent documents, and

by using said first list, said second list and said document IDs stored in said first document ID manager, a search result for said search request is generated.

5. The search system as set forth in claim 4, wherein in said generating of said search result,

said document IDs stored in said first document ID manager is removed from said first list to generate a third list;

and

a sum-set of said third list and said second list is generated.

6. A search system, comprising:

a first search index that is used to carry out an index search in response to a search request;

a second search index that is used to reflect first updated documents;

a first updating candidate document storage that stores data of second updated documents since a reflection processing of said first updated documents against said second search index began;

a second updating candidate document storage that stores data of said first updated documents used for said reflection processing against said second search index;

a first document ID manager, for added or modified document, that is used in search and stores document IDs of said first updated documents;

a first pattern matching document storage that stores said data of said first updated documents to carry out a string pattern matching in said search;

a second document ID manager, for added or modified document, that stores document IDs of said second updated documents since said reflection processing of said first updated documents against said second search index began;

a second pattern matching document storage that stores said data of said second updated documents since said reflection processing of said first updated documents against said second search index began, and

wherein, when said reflection processing of said first updated documents against said second search index is completed,

said second search index after said reflection processing is used as said first search index by switching,

said data of said second updated documents, which is stored in said first updating candidate document storage, is moved to said second updating candidate document storage by overwriting,

said document IDs of said second updated documents, which are stored in said second document ID manager, are moved to said first document ID manager by overwriting, and

said data of said second updated documents, which is stored in said second pattern matching document storage, is moved to said first pattern matching document storage by overwriting.

7. The search system as set forth in claim 6, wherein, before said reflection processing of said first updated documents against said second search index begins,

data of an updated document relating to a new update request is stored into said first and second updating candidate document storages,

a document ID of said update document relating to said new update request is stored into said first document ID manager, and

said data of said updated document relating to said new update request is stored in said first pattern matching document storage.

8. The search system as set forth in claim 7, wherein, after said reflection processing of said first updated document against said second search index began,

said first updating candidate document storage is cleared,

data of an updated document relating to a new update request is stored in said first updating candidate document storage,

said document ID of said updated document relating to said new update request is stored into said first and second document ID managers, and

said data of said updated document relating to said new request is stored in said first and second pattern matching document storages.

9. The search system as set forth in claim 6, wherein search relating to a search request is carried out by using said first search index to obtain a first list of document IDs of pertinent documents,

string pattern matching relating to said search request is carried out for said first pattern matching document storage to obtain a second list of document IDs of pertinent document, and

by using said first list, said second list and said document IDs stored in said first document ID manager, a search result for said search request is generated.

10. The search system as set forth in claim 6, wherein said updated document stored in said first and second updated candidate document storages is normalized for said search index, and

said data of said updated documents stored in said first and second pattern matching document storages is normalized for said string pattern matching.

11. A program embodied on a computer readable medium, for causing a computer to execute a search processing, said program comprising:

causing an index search unit for carrying out an index search using a search index before a document update to carry out said index search relating to a search request, and obtaining a first list of document IDs of pertinent documents;

causing a string pattern matching unit having document contents after said document update to carry out a string pattern matching relating to said search request for said document contents after said document update, and obtaining a second list of document IDs of pertinent documents; and

generating a search result for said search request by using said first and second lists and a third list of document IDs of documents relating to said document update.

12. A program embodied on a medium, for causing a computer to execute an updating process:

said computer comprising;

a first search index that is used to carry out an index search in response to a search request;

a second search index that is used to reflect first updated documents;

a first updating candidate document storage that stores data of second updated documents since a reflection processing of said first updated documents against said second search index began;

a second updating candidate document storage that stores data of said first updated documents used for said reflection processing against said second search index;

a first document ID manager, for added or modified document, that is used in search and stores document IDs of said first updated documents;

a first pattern matching document storage that stores said data of said first updated documents to carry out a string pattern matching in said search;

a second document ID manager, for added or modified document, that stores document IDs of said second updated documents since said reflection processing of said first updated documents against said second search index began;

a second pattern matching document storage that stores said data of said second updated documents since said reflection processing of said first updated documents against said second search index began, and

said program comprising;

when said reflection processing of said first updated documents against said second search index is completed,

using said second search index after said reflection processing as said first search index by switching;

moving said data of said second updated documents, which is stored in said first updating candidate document storage, to said second updating candidate document storage by overwriting;

moving said document IDs of said second updated documents, which are stored in said second document ID manager, to said first document ID manager by overwriting, and

moving said data of said second updated documents, which is stored in said second pattern matching document storage, to said first pattern matching document storage by overwriting.