Implementing Hash Function without MOD?

[Shanin]

Is there a way to implement Hash Function without MOD function?
I cannot guess that there is another way.

[j0nas]

Of course there is... A hash function is just a way to generate a value (does not even have to be unique) out of some data. If the data is a string for instance, a very simple string hash function could be:

int hashstring(const char *str)
{
int hash;
for (hash = 0; *str; str++)
hash += *str;
return hash;
}

(The above code may not be the best string hash function, it is just an example.)

[DeepButi]

j0nas,
correct answer ... but useless ;) as a hash value is usually used to access a table with a not infinite length, so it must be converted to a value in the range 0,table_size-1 and here MOD comes as the best(?) method.

(Useless message also :eek: )

[j0nas]

j0nas,
correct answer ... but useless ;) as a hash value is usually used to access a table with a not infinite length, so it must be converted to a value in the range 0,table_size-1 and here MOD comes as the best(?) method.

(Useless message also :eek: )

True but not... My hash string example could be used as a key: instead of storing variable-sized strings, we store INTs. This has nothing to do with the actual hash table.

But sure, normally you do MOD because you want an array index.

[yuwaraj]

Hi,
let me clear first u wanted
location=( hash_value % 10) , hash_value is returned by any hash_function and 10 is ur max limit
is that right ?
Then u r just need to know the range of ur (array say) i.e Index in that array should not
cross array length so u can write ur won function to calculate it using other operators.

[Yves M]

You could eliminate the MOD by making the hash table size a power of 2 and then use bitwise and to get the index. As in:

index = hash_value & 0xFF // for 256 slots

BUT, this is not a good idea, since in general, you will get many more collisions than if the size of your hash table is a prime number. So the performance will degenerate pretty badly.