Comment by vaibhavsagar

Doing a hashmap lookup is already very much like using a bloom filter: for any addressing scheme, if an element exists at a particular location, then either the key exists or it's a false positive. If no element exists, then there's no way the key exists. Adding an extra bloom filter would then come down to trading off space usage against a different probability of false positives, in addition to an extra lookup, and I think one would have to contrive an extremely unlikely scenario for it to make sense.