US 10,706,354 B2Granted

Estimating cardinality selectivity utilizing artificial neural networks

AssigneeInternational Business Machines Corporation

InventorsVincent Corvinelli, Huaxin Liu, Mingbin Xu, Ziting Yu, Calisto P. Zuzarte

Life of this patent

Estimating the term…

Loading patent…

Abstract

A database query comprising predicates may be received. Each predicate may operate on database columns. The database query may be determined to comprise strict operators. An upper bound neural network may be defined for calculating an adjacent upper bound and a lower bound neural network may be defined for calculating an adjacent lower bound. The upper bound neural network and the lower bound neural network may be trained using a selected value from data of a database table associated with the database query to be executed through the upper bound neural network and the lower bound neural network. The upper bound neural network and the lower bound neural network may be adjusted by passing in an expected value using an error found in expressions. The adjacent lower bound and the adjacent upper bound may be calculated in response to completion of initial training for the database columns.

Claims13

Dependent claims

A computer system for an adjacent value estimator for cardinality estimation using an artificial neural network, comprising:

one or more processors, one or more computer-readable memories, one or more computer-readable tangible storage medium, and program instructions stored on at least one of the one or more tangible storage medium for execution by at least one of the one or more processors via at least one of the one or more memories, wherein the computer system is capable of performing a method comprising:
receiving a database query comprising one or more predicates, wherein each predicate in the one or more predicates operates on one or more database columns of a database;
determining the received database query comprises one or more strict operators;
defining an upper bound neural network for calculating an adjacent upper bound and a lower bound neural network for calculating an adjacent lower bound based on receiving a database query;
training the defined upper bound neural network and the defined lower bound neural network using a selected value from a plurality of data within a database table associated with the received database query to be executed through the defined upper bound neural network and the defined lower bound neural network;
adjusting the defined upper bound neural network and the defined lower bound neural network by inputting an expected value using an error found in one or more expressions based on training the defined upper bound neural network and the defined lower bound neural network;
calculating the adjacent lower bound and the adjacent upper bound for a particular database column based on adjusting the defined upper bound neural network and the defined lower bound neural network;
inputting the calculated adjacent lower bound and calculated adjacent upper bound into the artificial neural network;
estimating a selectivity for the one or more database columns in response to inputting the calculated adjacent lower bound and calculated adjacent upper bound into the artificial neural network;
inputting the calculated adjacent lower bound and calculated adjacent upper bound in a query optimization process based on estimating the selectivity for the one or more database columns;
and
modifying the received database query with a plurality of equivalent non-strict operations for approximate query processing.

9 more elements

A computer program product for an adjacent value estimator for cardinality estimation using an artificial neural network, comprising:

one or more computer-readable storage medium and program instructions stored on at least one of the one or more tangible storage medium, the program instructions executable by a processor, the program instructions comprising:
program instructions to receive a database query comprising one or more predicates, wherein each predicate in the one or more predicates operates on one or more database columns of a database;
program instructions to determine the received database query comprises one or more strict operators;
program instructions to define an upper bound neural network for calculating an adjacent upper bound and a lower bound neural network for calculating an adjacent lower bound based on receiving a database query;
program instructions to train the defined upper bound neural network and the defined lower bound neural network using a selected value from a plurality of data within a database table associated with the received database query to be executed through the defined upper bound neural network and the defined lower bound neural network;
program instructions to adjust the defined upper bound neural network and the defined lower bound neural network by inputting an expected value using an error found in one or more expressions based on training the defined upper bound neural network and the defined lower bound neural network;
program instructions to calculate the adjacent lower bound and the adjacent upper bound for a particular database column based on adjusting the defined upper bound neural network and the defined lower bound neural network;
program instructions to input the calculated adjacent lower bound and calculated adjacent upper bound into the artificial neural network;
program instructions to estimate a selectivity for the one or more database columns in response to inputting the calculated adjacent lower bound and calculated adjacent upper bound into the artificial neural network;
program instructions to input the calculated adjacent lower bound and calculated adjacent upper bound in a query optimization process based on estimating the selectivity for the one or more database columns;
and
program instructions to modify the received database query with a plurality of equivalent non-strict operations for approximate query processing.

9 more elements

Dependencies are inferred from the claim text.