Poll of the Day > Here's a really difficult math/cryptography question

(+) Yeah! (+)

In my software I have a huge structure of data, maybe upwards of a GB. I eventually have to check the hash/checksum of this data, by reading the entire thing over again. However; what if I want to not have to re-read the entire thing to get that hash? What if I just want to update one item and update the hash accordingly?

Just writing this I came up with some idea on how to do this, but I'll see if anyone can think of a better solution.

Here is the hash method I'm using, btw. It's called a Knuth hash. For those who don't read code, I'm adding the value to 3,074,457,345,618,258,791, and then multiplying that value by 3,074,457,345,618,258,791, for each byte in the data.

ulong hashedValue = 3074457345618258791ul;
for (int i = 0; i < span.Length; i++) {
hashedValue += span[i];
hashedValue *= 3074457345618258799ul;
}

Solution below, don't read, figure it out!

Anyway, the solution I had come up would be to keep Knuth hashes of each segment of the data, some manageable amount. That way I can update a Knuth hash for one segment, and then Knuth hash those together to get the final hash. I was honestly hoping for a more fancy solution.

I have an unfair advantage because I work with these things all the time.

Fun fact

03074457345618258791 is the largest prime number that will fit in the long max value
18446744073709551615

I found an old list replacement class I wrote 3 years ago that I can upgrade with this functionality
https://github.com/jamieyello/NoDb/blob/master/SlothSerializer/Internal/LowMemList.cs

I knew it would be useful one day. ( 0:50 )
https://youtu.be/mGGFBycv2Hs?si=6e63EP4ZjbjZTH7N&t=50