Automatic search query generation and results set management
AssigneeTeradata US, Inc.
InventorsBrian J. Wasserman, George R. Hood, Thomas K. Ryan
Life of this patent
Estimating the term…
Abstract
Techniques are presented for automatic search query generation and results set management. A searcher makes selections from metadata; the metadata is associated with a particular lexicon or industry. Furthermore, the selections are mapped to fields associated with a data source. A thread-safe or unique search query is generated in response to the fields. The thread-safe search query is in a query language format which can query the data source. In an embodiment, the thread-safe search query is executed against the data source and the results returned are customized to rank the results, to sort the results, to filter the results, and/or to acquire a random sampling of the results.
Claims13
A method implemented in a computer-readable storage medium and configured to be processed by a computer, comprising:
- presenting metadata to a searcher, wherein the metadata is associated with a lexicon for a particular industry;
- receiving selections of the metadata from the searcher, wherein the metadata includes a topology and synonyms and custom defined relationships between terms and phrases associated with a particular industry, and wherein the metadata includes search operands to use with the terms and phrases;
- generating a thread-safe database query by using the selections as input parameters and generating the thread-safe database query as an Structured Query Language (SQL) query, wherein in building the thread-safe database query a select clause, a from clause, a where clause, and a group by clause are automatically generated, and wherein the thread-safe database query performs its data processing within its own thread of control and an execution space of the thread-safe data store query is unique and does not collide with other instances of database queries that concurrently process with the thread-safe database query, the input parameters including a login name for the searcher, matter identifier, a data source reference, a name for the thread-safe database query, and a name for metadata associated with the searcher;
- mapping the selections to fields of a database table;
- and
- populating the database query with the fields in a query language format of the database table in a results set table that is validated before it is presented to the searcher to ensure that query columns used in the search exist in the results set table.
A method implemented in a computer-readable storage medium and configured to be processed by a computer, comprising:
- interacting with a searcher to acquire metadata selections for a search query, wherein the metadata selections include a topology and synonyms and custom defined relationships between terms and phrases associated with a particular industry, and wherein the metadata selections include search operands to use with the terms and phrases, the metadata selections including a login name for the searcher, matter identifier, a data source reference, a name for the thread-safe database query, and a name for metadata associated with the searcher;
- mapping the metadata selections to fields associated with a database;
- automatically generating a unique search in response to the fields, wherein the unique search is generated to be a thread safe search and the unique search is a Structured Query Language (SQL) search, and wherein in generating the unique search a select clause, a from clause, a where clause, and a group by clause are automatically generated;
- querying the database with the unique search, and wherein the unique search performs its data processing within its own thread of control and an execution space of the unique search is unique and does not collide with other instances of database queries that concurrently process with the unique search;
- and
- presenting results associated with querying to the searcher in a results set table that is validated before it is presented to the searcher to ensure that query columns used in the unique search exists in the results set table.
Dependencies are inferred from the claim text.