Normally, a data set should be allocated for every possible PROJECT.GROUP.TYPE combination in the hierarchy. However, if the intent is to develop code in several hierarchies that merge in one main hierarchy, there might be no need to allocate some data sets. Allocating only the data sets that are actually used saves time when creating the hierarchy and minimizes DASD use and catalog entries. See Figure 1 for an example of a hierarchy that does not have all data sets allocated.

Only those data sets actually used in the hierarchy must be physically allocated. SCLM functions will execute successfully for hierarchies that contain unallocated data sets, as long as the unallocated data sets are not used. If a data set is not allocated and SCLM attempts to use the data set, an error message is issued.

Data sets can be added at any time. If you leave a data set unallocated and later find you need it, simply allocate the data set then.