Hola
Tengo que hacer una exposicion donde me hablan de este tema, la información que tengo al respecto es la siguiente:
Semijoins can still be beneficial with fast networks if they have very good selectivity an are implemented with bit arrays. A bit array B A[1:n] is useful in encoding the join attribute values present in one relation. Let us consider the semijoin R |X S. Then B A[i] is set to 1 if there exists a join attribute value A = val in relation S such that h(val)=i, where h is a hash function. Otherwise, B A[i] is set to 0. Such a bit array is much smaller than list of join attribute values. Therefore, transferring the bit array instead of the join attribute values to the site of relation R saves communication cost. The semijoin can be completed as follows. Each tuple of relation R, whose join attribute value is val, belongs to the semijoin if B A[h(val)]=1
Pero no entiendo muy bien el funcionamiento de los bit array sobre todo la parte de A=val y la busqueda con la funcion hash.
Agradeceria mucho si alguien entendiera un poco mas y me pudiera ayudar a entender ese parrafito que me ha causado dolores de cabeza
