A common phenomenon of time-qualified data are temporal repetitions, i.e., the association of multiple time values with the same data. In order to deal with finite and infinite temporal repetitions in databases we must use compact representations. There have been many compact representations proposed, however, not all of them are equally efficient for query evaluation. In order to show it, we define a class of simple queries on compact representations. We compare a query evaluation time on our proposed multi-granular compact representation GSequences with a query evaluation time on single-granular compact representation PSets, based on periodical sets. We show experimentally how the performance of query evaluation can benefit from the compactness of a representation and from a special structure of GSequences.