Index Partitioning and Scope Checking
AssigneeApple Computer, Inc.
InventorJohn Martin Hoernkvist
Life of this patent
Estimating the term…
Abstract
Systems and methods for processing an index are described. Searches are scope checked more efficiently using a forward lookup process based on the size of the requested search scope. In addition, an index is partitioned into separate stores based on a search scope that is learned based on where the user commonly conducts searches. As an example, a separate store may be created for a user's home directory should the user be conducting most of his or her searches in that directory. In addition to limiting the size of the index, during retrieval, intelligent index partitioning avoids the need to scope check a common search location.
Claims16
A machine-implemented method of partitioning an inverted index, the method comprising:
- determining a search scope for a user based on a location in which a user conducts searches;
- storing an index referencing items occurring within the determined search scope in an index store separate from an index referencing items occurring outside the determined search scope;
- and
- retrieving items from the index store in response to a search conducted in the location corresponding to the determined search scope.
A machine-implemented method of scoping a result set of an index, the method comprising:
- receiving a search having a query and a requested scope;
- running the query against an index to produce a result set;
- determining that a number of candidate results in the result set exceeds a threshold
- for filtering the result set to the requested scope using a reverse lookup;
- and filtering the result set to the requested scope using a forward lookup of the requested scope location.
A machine-implemented method of processing an index to facilitate searching the index, the method comprising:
- receiving a search having a query and a requested scope;
- running the query against an index to produce a result set;
- tracking a number of candidate results in the result set, the requested scope of the search, and an amount of time expended to filter the result set to the requested scope;
- identifying requested scope locations meeting criteria for inefficient searching based on the tracking;
- and
- moving identified scope locations meeting criteria for inefficient searching to a separate index.
A machine-readable medium containing executable program instructions for causing a data processing system to perform a method of processing an index to facilitate searching the index, the method comprising:
- receiving a search having a query and a requested scope;
- running the query against an index to produce a result set;
- tracking a number of candidate results in the result set, the requested scope of the search, and an amount of time expended to filter the result set to the requested scope;
- identifying requested scope locations meeting criteria for inefficient searching based on the tracking;
- and
- moving identified scope locations meeting criteria for inefficient searching to a separate index.
A data processing system comprising:
- means for receiving a search having a query and a requested scope;
- means for running the query against an index to produce a result set;
- means for tracking a number of candidate results in the result set, the requested scope of the search, and an amount of time expended to filter the result set to the requested scope;
- means for identifying requested scope locations meeting criteria for inefficient searching based on the tracking;
- and
- means for moving identified scope locations meeting criteria for inefficient searching to a separate index.
Dependencies are inferred from the claim text.