Abstract: Techniques herein optimize sort-merge join method for a band join. In an embodiment, for a query comprising a query block specifying a join between a first table and a second table, a band join condition is detected between the first table and the second table. Once the band join condition in detected, an execution plan is generated and executed. The execution of the execution plan includes: for a first row of at least a subset of first sorted rows, scanning second rows from a set of second sorted rows, joining each of said second rows with said first row, and ceasing to scan when encountering a row from the second sorted rows that falls outside a bound of said band join condition. Techniques also include parallelizing a workload by overlapping the distribution of rows to the same slave process and computing cost and cardinality estimation for enhanced band join.

Abstract: Techniques are described herein for supporting efficient evaluation of pivot operations. In an embodiment, a pivoting query on a data table is received. The query identifies pivot columns, measure columns, and grouping columns. In response to receiving the query, a staging table that comprises new pivot columns, aggregated columns, and new grouping columns may be created. For each particular unique combination of values of the grouping columns and the pivot columns in the data table, stored values in the measure columns in the data table are aggregated for certain rows in the plurality of rows in the data table. Each row of the certain rows comprises the each particular unique combination of values of the grouping columns and the pivot columns in the data table. Results of aggregating are stored in the staging table. Data that has been populated in the staging table is transposed into a result set.

Abstract: Techniques are described herein for supporting efficient evaluation of pivot operations. In an embodiment, a pivoting query on a data table is received. The query identifies pivot columns, measure columns, and grouping columns. In response to receiving the query, a staging table that comprises new pivot columns, aggregated columns, and new grouping columns may be created. For each particular unique combination of values of the grouping columns and the pivot columns in the data table, stored values in the measure columns in the data table are aggregated for certain rows in the plurality of rows in the data table. Each row of the certain rows comprises the each particular unique combination of values of the grouping columns and the pivot columns in the data table. Results of aggregating are stored in the staging table. Data that has been populated in the staging table is transposed into a result set.

Abstract: Query rewrite techniques are provided for rewriting an aggregate query with multiple groupings using multiple summary tables. The rewrite is accomplished by expanding the aggregate query into an equivalent UNION ALL query with multiple subqueries, each containing a single grouping. One or more of the subqueries is then rewritten to reference a summary table. Thus, a grouping of an aggregate query is derived by executing a subquery that either scans the summary table for the grouping, like in the exact-match mode, or by performing a roll-up operation, like in the roll-up mode. In addition, if the grouping of some of the subqueries cannot be derived from a summary table, subqueries are combined into another subquery that computes groupings from the base tables of the aggregate query.

Abstract: A method and apparatus for densification of data is provided. A method and apparatus for including a construct for partitioning data within a DML statement is provided. The construct for partitioning data may not necessarily perform a densification of data, and the densification of data may not necessarily include a construct for partitioning data. In an embodiment, the syntax of an OUTER JOIN is extended to include a PARTITION BY construct that may be used for data densification.

Abstract: A method and apparatus for densification of data is provided. A method and apparatus for including a construct for partitioning data within a DML statement is provided. The construct for partitioning data may not necessarily perform a densification of data, and the densification of data may not necessarily include a construct for partitioning data. In an embodiment, the syntax of an OUTER JOIN is extended to include a PARTITION BY construct that may be used for data densification.