Sources of data uncertainty and imprecision are numerous. A way of handle this uncertainty is to associate to data probabilistic annotations. Many such probabilistic database models have been proposed, both in the relational and semi-structured framework. The latter is particularly adapted to the management of uncertain data coming from automatic processes. An important problem, in the framework of probabilistic XML databases, is that of aggregation queries (count, sum, avg, etc.), that has not been studied to this day. In a model unifying the various semi-structured probabilistic models studied up to now, we present algorithms to compute the distribution of the aggregation (exploiting some regularity properties of the aggregation functions), and moments (especially, expectation and variance) of this distribution. We also prove the intractability of some of these problems.