After creating the distribution map, Adaptive Server employs two kinds of worker processes to perform different parts of the sort. These worker processes are called producer processes and consumer processes:
Producer processes read data from the input table and use the distribution map to determine the range to which each key value belongs. The producers distribute the data by copying it to the sort buffers belonging to the correct range.
Each consumer process reads the data from a range of the sort buffers and sorts it into subindexes, as described in “Range sorting”.
In Figure 9-1, two producer processes read data from the input table. Each producer process scans one table partition and distributes the data into ranges using the distribution map. For example, the first producer process reads data values 7, 2, 4, 5, and 9. Based on the information in the distribution map, the process distributes values 2 and 4 to the first consumer process, and values 7, 5, and 9 to the second consumer process.
