The Tetris-Algorithm for Sorted Reading from UB-Tree

1.1 Addresses, Areas and Regions We iteratively define an area A as a special subspace of a d-dimensional cube as follows: Split the cube with respect to every dimension in the middle, resulting in 2 subcubes numbered in some arbitrary but fixed order (for our implementation and this paper we used Z-ordering) from 1 to 2. An area A1 of level 1 consists of the first i1 closed subcubes. i1 determines A1 uniquely. We call i1 the address of A1 and write A1 = area(i1). The empty area has the address ε. To enlarge an area, we iteratively add an area with address i2 ∈{0,1,...,2-1} of the next subcube with number i1+1. The address of this enlarged area A2 is i1.i2, which is lexicographically larger than the address i1 of area A1. Next we may enlarge A2 by adding an area of the brother subcube i2+1 of i2, etc. The left part of figure 1 shows four areas area(0.0.1), area(1.3.2), area(2.1) and area(3) of a twodimensional universe. The shaded subcubes of the two-dimensional universe belong to the corresponding area.