I have some processes that are best performed using hash objects, often ending with use of the OUTPUT method to write results to a data set. But some of my datasets are so large that the correspondng hash objects need too much memory to run.
However, many of these tasks can be done in BY groups. Conceptually thiis would suggest using the hash object for one by-group, outputitng the results to a data set, then clearing the hash object, and doing the same process all over again for the next by group, followed by outputing it to the same data set.
But the OUTPUT method would only overwrite the data set. So I am asking for an APPEND method, which would work similarly to PROC APPEND. This would allow object-oriented processing of many large data sets with a much smaller demand on memory resources.