Microsoft Technology Licensing, LLC

Google Inc.

Campbell Bryce Fraser

Nigel Robin Ellis

Rodger N Kline

Srikanth Avadhanam

Peter Zabback

Ian Jose

Andrew Scott Richardson

Torsten Walther Grabs

Cesar Alejandro Galindo-Legaria

Milind Madhukar Joshi

Wei Yu

Vasileios Papadimos

Eugene Brevdo

Ciprian G Clinciu

Cesar A Galindo-Legaría

Lyric Pankaj Doshi

A method includes receiving a database query requesting a database to conditionally return one or more data blocks. The database is stored on memory hardware in communication with the data processing hardware and the database query includes a plurality of parameters characterizing the database query. The method includes generating a set of query plans. Each query plan in the set of query plans is configured to execute the database query using a different order of operations. The method includes training a model using historical database queries and generating, using the trained model, a query plan score for each query plan in the set of query plans. The method includes selecting, using the query plan score of each query plan in the set of query plans, a query plan from the set of query plans. The method also includes executing the database query using the selected query plan.

Database query optimization via parameter-sensitive plan selection

Described is a query optimizer comprising a query tuner that performs actual execution of query fragments to obtain actual results during compilation time, and uses those actual results to select a query plan. The actual results may be combined with estimates for fragments that were not executed. The tree may be traversed in a top-down traversal, processing every node. Alternatively, the tree may be traversed in a bottom-up traversal, re-deriving data for higher nodes as each lower level is completed. A limit, such as a time limit or level limit, may be used to control how much time is taken to determine the execution plan.

Query execution plans by compilation-time execution

Systems and methods are described that generate an execution plan for a query in a relational database system. The systems and methods generate the execution plan by generating one or more initial logical representations of the query, performing an exploration process around each of the one or more initial logical representations of the query, the performing of the exploration process around a particular initial logical representation of the query comprising applying transformation rules to generate one or more additional logical representations of the query that are logically equivalent to the particular initial logical representation of the query and that are within a maximum allowable transformation distance of the particular initial logical representation of the query, generating one or more execution plans for each initial logical representation of the query and each additional logical representation of the query, and selecting an execution plan from among the generated execution plans.

Distance-based logical exploration in a relational database query optimizer

Histograms formed based on samples of a population, such as histograms created from random page-level samples of a data store, are intelligently scaled to histograms estimating distribution of the entire population of the data store. As an optional optimization, where a threshold number of duplicate samples are observed during page-level sampling, the number of distinct values in the overall population data is presumed to be the number of distinct values in the sample data. Also, during estimation of distinct values of an overall population, a 'Chao' estimator can optionally be utilized as a lower bound of the estimate. The resulting estimate is then used when scaling, which can take domain knowledge of the data being scaled into account in order to prevent scaled estimates from exceeding the limits of the domain. Also, a 'sum of the parts' mathematical relationship can be taken into account during scaling that the sum of the scaled distinct values for each bin of an estimate histogram should total an estimate for the total distinct values of the entire population.

Generating histograms of population data by scaling from sample data

A compressed trie has nodes including multiple character sub-strings. Such multiple character storage reduces the number of nodes in the trie, thereby reducing the amount of memory required for storing the trie and reducing the amount of time required to perform matching. Furthermore, in such a compressed trie, sub-strings are stored in a single character string. Each node references its corresponding sub-string by the sub-string's starting position and length in the character string. Multiple nodes may reference a single sub-string. Thus, referencing rather than storing sub-strings in corresponding nodes eliminates repetitive sub-string storage, thereby reducing the amount of memory required for storing the trie.

System and method for using a compressed trie to estimate like predicates

Growing community of inventors

Redmond, WA, United States of America

Campbell Bryce Fraser