INVERTED INDEX PROCESSING
First Claim
Patent Images
1. A machine implemented method of indexing, the method comprising:
- storing more frequently occurring terms in a first table of an inverted index, the first table optimized for updating;
storing less frequently occurring terms in a second table, the second table not optimized for updating;
posting an item in which a more frequently occurring term occurs in the first table;
copying a previously posted item from the first table to a postings table; and
updating the posted item to point to the previously posted item in the postings table.
2 Assignments
0 Petitions
Accused Products
Abstract
Systems and methods for processing an index are described. In one exemplary method, a 2-level term table and postings table is used to generate postings lists. The postings lists are optimally stored in a postings file. Update sets for an index are optimally processed to update a index to a storage medium using flushing.
-
Citations
13 Claims
-
1. A machine implemented method of indexing, the method comprising:
-
storing more frequently occurring terms in a first table of an inverted index, the first table optimized for updating; storing less frequently occurring terms in a second table, the second table not optimized for updating; posting an item in which a more frequently occurring term occurs in the first table; copying a previously posted item from the first table to a postings table; and updating the posted item to point to the previously posted item in the postings table. - View Dependent Claims (2, 3, 4, 5)
-
-
6. A machine-implemented method of indexing, the method comprising:
-
allocating a first space on a storage medium for storing a postings file, the postings file containing data representing at least one list of items containing a term; writing the data to the allocated space in reverse order; and allocating a second space on a storage medium for storing the postings file, when writing the data would cause the postings file to exceed the amount of space allocated thus far. - View Dependent Claims (7, 8, 9)
-
-
10. A machine-implemented method of improving indexing, the method comprising:
-
storing an inverted index on a storage medium, the inverted index mapping a term to an item containing the term; sorting strings in an update set of strings representing items containing the term in a first order, the first order by a relative location of the term as mapped in the stored inverted index; and inserting strings from the update set of strings into the stored inverted index in accordance with the first order. - View Dependent Claims (11, 12, 13)
-
Specification