Many-core algorithms for in-memory column store databases
AssigneeSap SE
InventorsJonathan Dees, Peter Sanders, Franz Faerber, Jochen Seidel
Life of this patent
Estimating the term…
Abstract
A pattern can be identified in at least part of a query whose definition is received in a query request. The identified pattern can be matched with a set of pre-defined patterns, each of which has associated therewith at least one pre-compiled query execution sub-component of a plurality of pre-compiled query execution sub-components retained in a library. A plan for executing the query can be generated, for example by incorporating the pre-compiled query execution sub-component associated with the matched pattern into the plan based on a pseudo code representation of the plan derived from the definition.
Claims18
A non-transitory computer program product storing instructions that, when executed by at least one programmable processor forming part of at least one computing system, cause the at least one programmable processor to perform operations comprising:
- receiving a query request comprising a definition of a query of a database persisted in a column-based in-memory storage;
- identifying a pattern in at least part of the query;
- matching the identified pattern with a plurality of matching patterns among a set of pre-defined patterns, each of the pre-defined patterns having associated therewith at least one pre-compiled query execution sub-component of a plurality of pre-compiled query execution sub-components retained in a library, and wherein the matching includes selecting a selected matched pattern from among the plurality of matching patterns;
- selecting, based on the matching of the identified patterns with the plurality of matching-patterns, an optimal sequence for processing a plurality of tables that must be joined to respond to the query that avoids intermediate results and allows for the query to be performed in a single pass of the database, wherein at least one matching pattern other than the selected matched pattern produces intermediate results and requires more than two passes of the database;
- generating a plan for executing the query, the generating comprising incorporating the pre-compiled query execution sub-component associated with the selected matched pattern into the plan based on a pseudo code representation of the plan derived from the definition;
- and
- executing the query using the generated plan;
- and
- configuring the plan to include the optimal sequence for processing the plurality of tables;
- wherein the pseudo code is pre-prepared, pre-compiled, and stored in a library.
A system comprising:
- a column-based database storing data in-memory;
- at least one programmable processor;
- and
- a machine-readable medium storing instructions that, when executed by the at least one processor, cause the at least one programmable processor to perform operations comprising:
- receiving a query request comprising a definition of a query of the database;
- identifying a pattern in at least part of the query;
- matching the identified pattern with a plurality of matching patterns among a set of pre-defined patterns, each of the pre-defined patterns having associated therewith at least one pre-compiled query execution sub-component of a plurality of pre-compiled query execution sub-components retained in a library, and wherein the matching includes selecting a selected matched pattern from among the plurality of matching patterns;
- selecting, based on the matching of the identified patterns with the plurality of matching-patterns, an optimal sequence for processing a plurality of tables that must be joined to respond to the query that avoids intermediate results and allows for the query to be performed in a single pass of the database, wherein at least one matching pattern other than the selected matched pattern produces intermediate results and requires more than two passes of the database;
- generating a plan for executing the query, the generating comprising incorporating the pre-compiled query execution sub-component associated with the selected matched pattern into the plan based on a pseudo code representation of the plan derived from the definition;
- and
- executing the query using the generated plan;
- configuring the plan to include the optimal sequence for processing the plurality of tables;
- wherein the pseudo code is pre-prepared, pre-compiled, and stored in a library.
A computer-implemented method comprising:
- receiving a query request comprising a definition of a query of a database persisted in a column-based in-memory storage, the query comprising a plurality of subparts;
- identifying, for each of at least two subparts of the query, a pattern from at least part of the query;
- matching each identified pattern with a pattern among a set of pre-defined patterns, each of the pre-defined patterns having associated therewith at least one pre-compiled query execution sub-component of a plurality of pre-compiled query execution sub-components retained in a library, and wherein the matching includes selecting a selected matched pattern from among the plurality of matching patterns
- selecting, based on the matching of the identified patterns with the plurality of matching-patterns, an optimal sequence for processing a plurality of tables that must be joined to respond to the query that avoids intermediate results and allows for the query to be performed in a single pass of the database, wherein at least one matching pattern other than the selected matched pattern produces intermediate results and requires more than two passes of the database;
- generating a plan for executing the query, the generating comprising incorporating the pre-compiled query execution sub-component associated with the selected matched pattern into the plan based on a pseudo code representation of the plan derived from the definition;
- and
- executing the query using the generated plan;
- and
- configuring the plan to include the optimal sequence for processing the plurality of tables;
- wherein the pseudo code is pre-prepared, pre-compiled, and stored in a library.
Dependencies are inferred from the claim text.