Hello, Antei, you wrote: I>>> Hello, samius, you wrote: A> Thoughts aloud... A> And that if to take two algorithm and: A> - on the basis of input 5-7 string columns to count two and A> - them, i.e. as a result we receive A> HashFunction1 (col1, col2..., col7) + delimiter + HashFunction2 (col1, col2..., col7) A> It should reduce probability practically to 0. A> That think (practically, not )? Filter Bluma detected https://ru.wikipedia.org/wiki/%D0%A4%D0 … 0%BC%D0%B0 Filtr Bluma represents bit map from m bit. Initially, when the data structure stores empty set, all m the bit is nullified. The user should define k independent hash functions h1, ..., hk, displaying each element in one of m bit map positions in uniform enough image. For element adding e it is necessary to write down units on each of positions h1 (e), ..., hk (e) bit map. For check of an accessory of an element e to set of storable elements, it is necessary to check up a state of bits h1 (e), ..., hk (e). If at least one of them is equal to zero, the element cannot belong to set (differently at its adding all these bits would be installed). If all of them are equal to unit the data structure informs that belongs to set. Thus there can be two situations: or the element really belongs to set, or all these bits appeared are installed on randomness at adding of other elements, as is a source of false actuatings in this data structure.