Spatial joins in multi-processing computing systems including massively parallel processing database systems

A computer-implemented method of processing spatial joins of data in a parallel processing computing system, wherein the parallel processing computing system includes one or more processors configured to provide multiple processing units configured to process data in parallel, and wherein the computer-implemented method comprises:

• obtaining first and second bounding data sets respectively for first and second data sets of a database, wherein the first and second bounding data sets of the database respectively include first and second bounding information, and wherein the first and second bounding information respectively include representation of bounds or limits of the first and second datasets in a space that a spatial join can be performed between the first and second data sets to effectively join at least a respective part of at one of the first and second data sets spatially in the space;
• determining, based on the obtained first and second bounding data sets, a spatial join bounding space, as an intersection of the first and second bounding data sets, in the space where the first and second data sets are to be spatially joined;
• partitioning the determined spatial join bounding space of the first and second data sets are to be spatially joined into multiple sub-partitions of the spatial join bounding space;
• and
• assigning two or more of the multiple sub-partitions of the spatial join bounding space respectively to two or more processing units of a multi-processing computing system for processing the sub-partitions in parallel, wherein the multi-processing computing system includes at least one physical processor configured to provide multiple processing units to process data in parallel, thereby allowing the two or more processing units of a multi-processing computing system to process in parallel their respective assigned sub-partitions of the spatial join bounding space.

A computer-implemented method of processing a spatial join in a parallel processing computing system, wherein the parallel processing computing system includes one or more processors configured to provide multiple processing units configured to process data in parallel, and wherein the computer-implemented method comprises:

• assigning each sub-partition of a join grid to a processing unit of a multi-processing computing system for processing the sub-partitions in parallel;
• distribute each one of the join pairs based on a distribution strategy;
• perform an evaluation of each one of the distributed join pairs on the distribution strategy;
• and
• generate the join results at least partly based on one or more of the following:
• one or more join predicates, and one or more residual join conditions.

A non-transitory computer readable storage medium storing at least executable computer code for processing spatial joins of data in a parallel processing computing system, wherein the parallel processing computing system includes one or more processors configured to provide multiple processing units configured to process data in parallel, and wherein when the executable computer code is executed it performs at least the following:

• obtain first and second bounding data sets respectively for first and second data sets of a database, wherein the first and second bounding data sets of the database respectively include first and second bounding information, wherein the first and second bounding information respectively include representation of bounds or limits of the first and second datasets in a space that a spatial join can be performed between the first and second data sets to effectively join at least a respective part of at one of the first and second data sets spatially in the space;
• determine, based on the obtained first and second bounding data sets, a spatial join bounding space, as an intersection of the first and second bounding data sets, in the space where the first and second data sets are to be spatially joined;
• partition the determined spatial join bounding space of the first and second data sets are to be spatially joined into multiple sub-partitions of the spatial join bounding space;
• and
• assign two or more of the multiple sub-partitions of the spatial join bounding space respectively to two or more processing units of a multi-processing computing system for processing the sub-partitions in parallel, wherein the multi-processing computing system includes at least one physical processor configured to provide multiple processing units to process data in parallel, thereby allowing the two or more processing units of a multi-processing computing system to process in parallel their respective assigned sub-partitions of the spatial join bounding space.