US Patent Application 17735720. QUERY PROCESSING BASED ON STOCHASTIC PREDICTION MODEL simplified abstract
Contents
QUERY PROCESSING BASED ON STOCHASTIC PREDICTION MODEL
Organization Name
Inventor(s)
QUERY PROCESSING BASED ON STOCHASTIC PREDICTION MODEL - A simplified explanation of the abstract
- This abstract for appeared for US patent application number 17735720 Titled 'QUERY PROCESSING BASED ON STOCHASTIC PREDICTION MODEL'
Simplified Explanation
This abstract describes a method for query processing in a database using a stochastic prediction model. The method involves maintaining information about past query plans and building a data structure that represents the probabilities of transitioning between different database operators. When a new query is received, a stochastic prediction algorithm is applied to identify one or more candidate query plans based on the data structure. These candidate query plans are then executed to retrieve the required data from the database for the new query.
Original Abstract Submitted
Methods, apparatus, and processor-readable storage media for query processing based on a stochastic prediction model are provided herein. An example computer-implemented method includes maintaining information associated with a set of database operators corresponding to one or more historical query plans, wherein each of the historical query plans is associated with processing of a corresponding historical query in a database; building, based on the maintained information, a data structure comprising probabilities of each database operator in the set transitioning to each of one or more of the other database operators in the set; obtaining a new query to be processed in the database; applying a stochastic prediction algorithm to identify, from among the historical query plans, at least one candidate query plan for the new query based on the data structure; and executing the at least one candidate query plan to access data from the database for the new query.