Abstract: Described is an approach for performing an efficient processing of table functions/polymorphic table functions. The approach does not create additional cursors for processing a table or polymorphic table function invoked within a SQL statement. Instead, the approach handles the SQL statement that invokes the table/polymorphic table function as a single cursor.

Abstract: In an artificial neural network, integrality refers to the degree to which a neuron generates, for a given set of inputs, outputs that are near the border of the output range of a neuron. From each neural network of a pool of trained neural networks, a group of neurons with a higher integrality is selected to form a neural network tunnel (“tunnel”). The tunnel must include all input neurons and output neurons from the neural network, and some of the hidden neurons. Tunnels generated from each neural network in a pool are merged to form another neural network. The new network may then be trained.

Abstract: Disclosed is an approach to define non-SQL functions as a SQL macro, which is then processed inline with SQL statements. By inlining the macro into SQL, this approach avoids cross-engine context switches as well as making it possible for a SQL query optimizer to generate better plans that execute more efficiently.

Abstract: Described is an approach for performing an efficient processing of table functions/polymorphic table functions. The approach does not create additional cursors for processing a table or polymorphic table function invoked within a SQL statement. Instead, the approach handles the SQL statement that invokes the table/polymorphic table function as a single cursor.

Abstract: Techniques are provided for caching external data sources for SQL processing. According to an embodiment, a database system receives a query that requires foreign data from an external data source. In response to receiving the query, the database system determines whether the foreign data currently resides in a cache of the database system. In response to determining that the foreign does not currently reside in the cache, the database system retrieves the foreign data from the external data source, maps the foreign data to a column granule of a table, and stores the column granule in the cache. In another embodiment, the database system receives a second query that requires the foreign data. In response to determining that the foreign data is cached with the stored column granule, the database system retrieves the column granule from the cache.

Abstract: Techniques are provided for caching external data sources for SQL processing. According to an embodiment, a database system receives a query that requires foreign data from an external data source. In response to receiving the query, the database system determines whether the foreign data currently resides in a cache of the database system. In response to determining that the foreign does not currently reside in the cache, the database system retrieves the foreign data from the external data source, maps the foreign data to a column granule of a table, and stores the column granule in the cache. In another embodiment, the database system receives a second query that requires the foreign data. In response to determining that the foreign data is cached with the stored column granule, the database system retrieves the column granule from the cache.

Abstract: Techniques are provided for caching external data sources for SQL processing. According to an embodiment, a database system receives a query that requires foreign data from an external data source. In response to receiving the query, the database system determines whether the foreign data currently resides in a cache of the database system. In response to determining that the foreign does not currently reside in the cache, the database system retrieves the foreign data from the external data source, maps the foreign data to a column granule of a table, and stores the column granule in the cache. In another embodiment, the database system receives a second query that requires the foreign data. In response to determining that the foreign data is cached with the stored column granule, the database system retrieves the column granule from the cache.

Abstract: Techniques are provided for caching external data sources for SQL processing. According to an embodiment, a database system receives a query that requires foreign data from an external data source. In response to receiving the query, the database system determines whether the foreign data currently resides in a cache of the database system. In response to determining that the foreign does not currently reside in the cache, the database system retrieves the foreign data from the external data source, maps the foreign data to a column granule of a table, and stores the column granule in the cache. In another embodiment, the database system receives a second query that requires the foreign data. In response to determining that the foreign data is cached with the stored column granule, the database system retrieves the column granule from the cache.

Abstract: A method and apparatus for selectively allowing data presented in a join view to be modified is disclosed. Data presented in a join view is allowed to be modified if provided to the join view by a key-preserved table. A table is a key-preserved table with respect to a join view if each of the keys of the table is a key of the join view. Key-preserved tables may be identified by inspecting the definition of a join view to identify the base tables of the join view and to generate a graph. In the graph, each node corresponds to one of the base tables. Directed edges between nodes of the graph correspond to relationships, specified in the definition of the join view, between the base tables. The graph is reduced to strong components, each of which corresponds to a set of the base tables in which each base table in the set is routed to each other base table in the set by one or more directed edges.

Abstract: A method for transforming database queries that have EXISTS-equivalent subqueries (i.e. subqueries that are semantically equivalent to EXISTS or NOT EXISTS subqueries) is disclosed. Logical expressions are removed from a WHERE CLAUSE in the original EXISTS-equivalent subquery to produce a revised subquery. For each of the logical expressions that are removed, a substitute query component is generated based on whether (1) the logical expression contains any reference to any table in a FROM-list in the EXISTS-equivalent subquery and (2) the original EXISTS-equivalent subquery is equivalent to an EXISTS or a NOT EXISTS subquery. The transformed query is created by replacing the original EXISTS-equivalent subquery with the revised subquery and the query components that were generated for each of removed logical expressions. Some of the substitute query components, including the revised subquery, may be merged based on knowledge of access paths.