Method for mapping an index of a database into an array of files

• US 5,787,435 A
• Filed: 08/09/1996
• Issued: 07/28/1998
• Est. Priority Date: 08/09/1996
• Status: Expired due to Term

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Digital Equipment Corporation (HP Inc.)
• Inventors: Burrows, Michael
• Primary Examiner(s): AMSBURY, WAYNE P
• Application Number:
  US08/700,748
• Time in Patent Office: 718 Days
• Field of Search: 395/613, 395/610, 395/603, 395/615, 395/605, 395/602, 707/102, 707/5, 707/3, 707/2
• US Class Current: 1/1
• CPC Class Codes: G06F 16/319 Inverted lists
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99937 Sorting
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

A computer implemented method maps an index of information of a database into an array of files. The information is stored as records of a database. Batches of records are parsed into words and locations, each word representing a portion of the information of a particular record, the locations are sequentially assigned to the words in the order of parsing the words. An index entry is generated for each unique word, each index entry including the unique word and all of the locations where the unique word occurs in the database. Each unique word is hashed to determine a bucket number. Index entries having identical bucket numbers are written to a single index file associated with the bucket number in a collating order of the unique words. A summary file is generated for each index file by periodically sampling the data being written to the index file. Index files and their associated summary files are grouped into a tier of files, there being one tier for each batch of records parsed.

Claims

1. A computer implemented method for mapping an index of information of a database as a plurality of records, comprising:

parsing a batch of the records into words and locations, each word representing a portion of the information of a particular record, and each location assigned sequentially in the order of parsing the words;

generating an index entry for each unique word, each index entry including the unique word and all of the locations of the unique word;

hashing each unique word to determine a bucket number;

writing index entries having identical bucket numbers to a single index file in an order according to their unique words;

generating a summary file for each index file by periodically sampling the data being written to the index file;

grouping the index files and their associated summary files into a tier of files representing an index of the batch of records.

2. The method of claim 1 further comprising:

repeatedly performing the steps of claim 1 for a plurality of batches to produce a plurality of tiers of files.