Description of sup- and inf-preserving aggregation functions via families of clusters in data tables