Method for storing text annotations with associated type information in a structured data store

• US 7,558,786 B2
• Filed: 12/21/2005
• Issued: 07/07/2009
• Est. Priority Date: 12/24/2004
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Robson, Christine Margaret, Watanabe, Yuji
• Primary Examiner(s): Ali; Mohammad
• Assistant Examiner(s): Hocker; John P
• Application Number:
  US11/314,766
• Publication Number:
  US 20060143164A1
• Time in Patent Office: 1,294 Days
• Field of Search: 707 1- 10
• US Class Current: 1/1
• CPC Class Codes: G06F 16/2272 Management thereof
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

A database system, a computer executable method, a computer executable program for the database system, and a method for updating an indexing tables in a database system are provided. The database system comprises a database storing item tables of items appended by timestamps. Each of the items is further categorized by timestamp into the tables, and the item tables are indexed by indexing data. The database system further comprises an indexing table for the indexing data, listing the earliest timestamp of the item table together with the indexing data and a system manager for searching the item tables using the indexing table and a maintenance module for the indexing table.

Claims

1. A database search method in a database system, said database system comprising a computer including a CPU, a memory, and a storage device, said database being stored in said storage device, said database search method making said computer execute the steps of:

storing parts tables of items appended by timestamps in a database, including parts table identifiers, and storing items in an indexing table in which the update time and parts table identifier are stored, said parts tables being indexed by indexing data, said database and said parts tables being updatable by appending only,

categorizing said items by timestamp and parts table identifier into said parts tables, and indexing said parts tables by indexing data such that time differential between given items is correlated to time differential between indexing data corresponding to said given items,

appending said database with items from a second database when a predetermined number of new items becomes available in said second database,

storing an indexing table for said indexing data, listing a newest timestamp of said parts tables together with said indexing data,

searching for a target item from said indexing table among said items from said parts tables,

repairing said indexing table with respect to time evolution by a maintenance module, said maintenance module executing further steps of;

updating, in response to an event, said indexing table by appending said indexing table with items from said parts tables, and said updating occurs when a new one of said parts tables is issued,

monitoring correlation between said parts tables and said indexing table by comparing a timestamp of an item with a newest timestamp in said parts tables with a timestamp of an entry in said indexing table, said entry in said indexing table corresponding to said item with said newest timestamp in said parts tables, and

examining said correlation of said indexing table using said newest timestamp and said indexing data as examination keys,

repairing said indexing table by predicting an insert sequence of entries in said indexing table if a value of said newest timestamp of said parts table identifier is less than a value of said timestamp corresponding to the parts table identifier of said entry in said indexing table for said item having said newest timestamp,

returning a positive value, indicating said indexing table is in regular order if said value of said newest timestamp of said parts table identifier is greater than a value of said timestamp corresponding to the parts table identifier of said entry in said indexing table for said item having said newest timestamp,

indicating that said indexing table is in regular order if said value of said newest timestamp of said parts table identifier is between two values of timestamps corresponding to the parts table identifier of said indexing table,

repairing said indexing table by predicting the insert sequence of entries in the indexing table and inserting the insert sequence corresponding to the parts table identifier, if said value of said newest timestamp of said parts table identifier is between two values of timestamps of said indexing table, with the two values of timestamps of said indexing table corresponding to different parts table identifiers,

executing a binary search of said database with referring said indexing table,

determining whether or not an appending of records of said indexing table is required depending on search keys, and

depending on said determination step, if said appending is required, calling a repairing function so as to append said update records while executing said binary search.

2. The database search method of claim 1, said method further making said computer execute the step of appending records to said indexing table in response to results of said examination of said monitor tool according to said timestamp and said indexing data.

3. The database search method of claim 1, said method further including the step of executing a binary search as part of said repaired indexing table.

4. A method for updating an indexing tables in a database system, said database system comprising a computer including a CPU, a memory, and a storage device, said database being stored in said storage device, said method making said computer execute the steps of:

storing parts tables of items appended by timestamps in a database,

categorizing said items by timestamp and parts table identifier into said parts tables, and indexing said parts tables by indexing data such that time differential between given items is correlated to time differential between indexing data corresponding to said given items,

appending said database with items from a second database when a predetermined number of new items becomes available in said second database,

listing a newest timestamp of said parts tables together with said indexing data,

monitoring a correlation between said parts tables and said indexing table, wherein said correlation is responsive to a difference between timestamp value of an entry in said parts tables and timestamp value of an entry in said indexing table for a given parts table identifier,

invoking from said parts tables said newest timestamp together with said indexing data,

examining said correlation of said indexing table using said invoked newest timestamp and said indexing data as examination keys,

appending update records to said indexing table in response to results of said examination of said monitor tool according to said timestamp and said indexing data by determining record sequence and appending said update records in accordance with record sequence,

updating said indexing table by predicting an insert sequence of entries in said indexing table if a value of said newest timestamp of said parts table identifier is less than a value of said timestamp corresponding to the parts table identifier of said entry in said indexing table for said item having said newest timestamp, and

returning a positive value, indicating said indexing table is in regular order if said value of said newest timestamp of said parts table identifier is greater than a value of said timestamp corresponding to the parts table identifier of said entry in said indexing table for said item having said newest timestamp,

indicating that said indexing table is in regular order if said value of said newest timestamp of said parts table identifier is between two values of timestamps corresponding to the parts table identifier of said indexing table,

updating said indexing table by predicting the insert sequence of entries in the indexing table and inserting the insert sequence corresponding to the parts table identifier, if said value of said newest timestamp of said parts table identifier is between two values of timestamps of said indexing table, with the two values of timestamps of said indexing table corresponding to different parts table identifiers,

wherein said updating an indexing table is executed when a new one of said parts tables is issued,

executing a binary search of said database with referring said indexing table,

determining whether or not an appending of records of said indexing table is required depending on search keys, and

depending on said determination step, if said appending is required,

calling a repairing function so as to append said update records while executing said binary search.