> I feel like using a straight bloom filter will get you some false positives That is exactly the property of bloom filter searches (unless I'm misunderstandin

> I feel like using a straight bloom filter will get you some false positives That is exactly the property of bloom filter searches (unless I'm misunderstanding) - you should never get a false negative but can expect false positives. The design of the filter will define how many false positives are likely for a given lookup in a given dataset. > might be good for a first-pass. Exactly. If the number of false positives is low then you can perform an exhaustive search on the few results to exclude the wrong ones. You won't miss anything as there are no false negatives. Of course defining the bloom filter such that the number of false positives is low enough to "win" by a significant margin might not be easy, particularly for growing rather than static data, just like choosing an optimal hash function often isn't.