Reason behind hashcode for String

Question

I want to know how the formula was designed for the hashcode of String class.
Also as a continuation, is there a difference between the generation of hashcode in different versions of java (may be in other classes as well) ? :)

Answer 1

sabre150 wrote:-
public class TwoStringWithTheSameHashCode
public static void main(String[] args)
int hash1 = "ABCDEa123abc".hashCode();
int hash2 = "ABCDFB123abc".hashCode();
System.out.println(hash1 + " " + hash2);
Good point. I mean, it's not impossible to get collisions. I have to admit I didn't really sit and analyze the function to try to calculate the probability. So the answer may very well be that it tends to spread the hashes out in practice. Well, at least, that's the whole idea, right? So maybe there is a better algorithm for it.
As for full uniqueness, that's impossible in general -- that's the definition of a hashcode, after all (unless your hashcode is just to return the same value). There are, however, industry-standard algorithms that generate pretty large sets of possible hashcodes (MD5, SHA1, SHA256...)
In the end, the best hashcode depends on your data set and what you're willing to sacrifice. If you have a data set of only 10 distinct values, you're (theoretically) not going to benefit much from a hash much at all. If you have millions/billions/etc of unique values (say, all strings up to 1024 chars), then a hashcode which pushes it into 256 bytes is going to collide eventually if you put every possible string into it. The trick is much collision you're okay with, how big your target space is, and how much work you want to do to get a hash.