Automatic live-aggregate projection refresh responsive to on-demand flattened table refresh

A non-transitory computer-readable data storage medium storing program code executable by a database management system (DBMS) running on a computing system to:

• refresh on-demand a selected denormalized column of a flattened table (FT) of a database of the DBMS with respect to a selected partition storing row data for the selected denormalized column;
• and
• responsive to refreshing on-demand the selected denormalized column of the FT, automatically refresh a plurality of projections defined on the FT of the database and that include the selected denormalized column and automatically refresh a live-aggregate projection (LAP) defined on the FT of the database and that corresponds to an aggregate column of the FT defining an aggregation over the selected denormalized column,
• wherein each projection corresponds to a selected set of denormalized columns of the FT of the database,
• wherein automatic refresh of the LAP begins as soon as one of the projections has been automatically refreshed, the LAP automatically refreshed from the one of the projections that has been automatically refreshed,
• wherein the LAP physically materializes the aggregation of the selected normalized column over the selected denormalized column as defined by the aggregate column in that the LAP stores the aggregation of the selected normalized column over the selected denormalized column, redundant to storage of the selected normalized column and the selected denormalized column within a plurality of partitions including the selected partition,
• and wherein the LAP is coupled to the selected denormalized column of the FT, such that changes in the FT are automatically propagated to the LAP through an intervening projection including the selected denormalized column.

A computing system comprising:

• physical resources, including a storage device storing a database of a database management system (DBMS), a processor, and memory;
• and
• DBMS logic implemented using the physical resources to:
• maintain a flattened table (FT) of the database logically materializing a plurality of columns over a plurality of partitions on the storage device, the columns comprising a plurality of normalized columns, a plurality of denormalized columns from a plurality of source tables of the database, and an aggregate column defining an aggregation of a selected normalized column over a selected denormalized column;
• maintain a plurality of projections on the FT of the database, each projection corresponding to a selected set of the columns of the FT;
• maintain a live-aggregate project (LAP) on the FT of the database, the LAP corresponding to the aggregate column and physically materializing the aggregation of the selected denormalized column over the selected denormalized column as defined by the aggregate column;
• responsive to on-demand refreshing the selected denormalized column of the FT over a selected partition, automatically refresh the projections that include the selected denormalized column;
• and
• responsive to automatically refreshing the projections that include the selected denormalized column, automatically refresh the LAP as soon as one of the projections including the selected denormalized column has finished refreshing,
• wherein the LAP physically materializes the aggregation of the selected normalized column over the selected denormalized column as defined by the aggregate column in that the LAP stores the aggregation of the selected normalized column over the selected denormalized column, redundant to storage of the selected normalized column and the selected denormalized column within the partitions,
• and wherein the LAP is coupled to the selected denormalized column of the FT, such that changes in the FT are automatically propagated to the LAP through an intervening projection including the selected denormalized column.