Grouping in analytical databases

A computer program product for grouping records of a data table, the data table being stored on multiple data blocks, comprising a computer readable storage medium having computer readable program code embodied therewith, wherein the computer readable program code when executed on a computer causes the computer to:

• receive a query against the data table;
• determine that processing of the query involves grouping of records of the data table with respect to a grouping attribute;
• determine a set of distinct value ranges of the grouping attribute according to a scan list that comprises a scan order of the data blocks ordered based upon data value ranges of the data blocks, with each distinct value range of the grouping attribute comprising an associated scan position in the scan order indicative of a data block in the scan order after which no further data records within a respective distinct value range are encountered;
• and
• scan the data blocks in accordance with the scan list and process the query according to a query execution plan, wherein the processing of the query involves, in response to reaching a scan position:
• calculating an aggregate of scanned records corresponding to records of the data blocks of the respective distinct value range associated with the scan position;
• and
• flushing results relating to the respective distinct value range to subsequent operators in the query execution plan.

A computer system for grouping records of a data table, comprising:

• at least one memory that stores the data table on multiple data blocks;
• a network interface card;
• and
• at least one processor programmed to:
• receive, via the network interface card, a query against the data table;
• determine that processing of the query involves grouping of records of the data table with respect to a grouping attribute;
• determine a set of distinct value ranges of the grouping attribute according to a scan list that comprises a scan order of the data blocks ordered based upon data value ranges of the data blocks, with each distinct value range of the grouping attribute comprising an associated scan position in the scan order indicative of a data block in the scan order after which no further data records within a respective distinct value range are encountered;
• and
• scan the data blocks in accordance with the scan list and process the query according to a query execution plan, wherein the processing of the query involves, in response to reaching a scan position:
• calculating an aggregate of scanned records corresponding to records of the data blocks of the respective distinct value range associated with the scan position;
• and
• flushing results relating to the respective distinct value range to subsequent operators in the query execution plan.