Let exr(n;Kr+1r) be the maximum number of r-edges that can be placed on n vertices without forming a Kr+1r (the r-uniform complete graph on r+1 vertices).
Is every r-hypergraph G on n vertices the union of at most exr(n;Kr+1r) many copies of Krr and Kr+1r, no two of which share a Krr?