For an MDR program to complete it needs to process all available intervals. Interval Processing refers to the custom source code that processes the data relating to each interval – for example processes all Sales Orders with the keys 0000000001 – 0000001000. Note that the processing of each interval should be performed independently of the processing of any other interval. The main focus of the processing component of the MDR program is to process each interval independently; this means dealing with only with the master and transactional data within the range of the interval.
When an MDR program is executed, it is run with X jobs (where X is a number from 1 upwards). MDR achieves its performance gains by executing each job in parallel, hence the reason why the processing of one interval should be independent of any others. As each interval is processed, the results of that interval are stored. MDR provides a number of easy to use statements as a generic storage mechanism for each interval’s results. These will be explained in more detail in this document.