Hash Code Of A String

[j1mb0jay]

I am currently working on a dictionary populating program. I currently
have a socket connection my local news server and am trawling through
all of the articles looking for new words. I am currently using Java to
do this but would like to move the source to C#. Java's String class has
a method that hashes strings. I was wondering if C# has a method which
does the same?

In my Java version of the program I am using the Multiply Add and Divide
(MAD) method for the compression of the hash code, does C# have any
built in functions(metho ds) that will do this for me, or does anyone
know of a more efficient way?

j1mb0jay

[=?Utf-8?B?S0g=?=]

All .NET types have the GetHashCode() method, at minimum inherited from
System.Object.

string str = "fubar";

int hc = str.GetHashCode (); // 418978654
"j1mb0jay" wrote:

I am currently working on a dictionary populating program. I currently
have a socket connection my local news server and am trawling through
all of the articles looking for new words. I am currently using Java to
do this but would like to move the source to C#. Java's String class has
a method that hashes strings. I was wondering if C# has a method which
does the same?

In my Java version of the program I am using the Multiply Add and Divide
(MAD) method for the compression of the hash code, does C# have any
built in functions(metho ds) that will do this for me, or does anyone
know of a more efficient way?

j1mb0jay

[j1mb0jay]

KH wrote:

All .NET types have the GetHashCode() method, at minimum inherited from
System.Object.

string str = "fubar";

int hc = str.GetHashCode (); // 418978654
"j1mb0jay" wrote:

>I am currently working on a dictionary populating program. I currently
have a socket connection my local news server and am trawling through
all of the articles looking for new words. I am currently using Java to
do this but would like to move the source to C#. Java's String class has
a method that hashes strings. I was wondering if C# has a method which
does the same?

In my Java version of the program I am using the Multiply Add and Divide
(MAD) method for the compression of the hash code, does C# have any
built in functions(metho ds) that will do this for me, or does anyone
know of a more efficient way?

j1mb0jay

Much the same as Java the GetHashCode() methodology. How would you
convert the 418978654 to a number that I could use as an index into a
fixed size array?

j1mb0jay

[=?Utf-8?B?S0g=?=]

That's really a whole different question; a couple options:

1) GetHashCode() returns a 32-bit interger, so if you created an array of
size UInt32.MaxValue you could use the hashcode directly

2) You could % the hashcode by the size of your array

Neither of those are actually viable options of course. What are you
actually trying to accomplish?
"j1mb0jay" wrote:

KH wrote:

All .NET types have the GetHashCode() method, at minimum inherited from
System.Object.

string str = "fubar";

int hc = str.GetHashCode (); // 418978654
"j1mb0jay" wrote:

I am currently working on a dictionary populating program. I currently
have a socket connection my local news server and am trawling through
all of the articles looking for new words. I am currently using Java to
do this but would like to move the source to C#. Java's String class has
a method that hashes strings. I was wondering if C# has a method which
does the same?

In my Java version of the program I am using the Multiply Add and Divide
(MAD) method for the compression of the hash code, does C# have any
built in functions(metho ds) that will do this for me, or does anyone
know of a more efficient way?

j1mb0jay

Much the same as Java the GetHashCode() methodology. How would you
convert the 418978654 to a number that I could use as an index into a
fixed size array?

j1mb0jay

[Peter Duniho]

On Fri, 11 Jan 2008 15:24:07 -0800, j1mb0jay <no**@none.comw rote:

Much the same as Java the GetHashCode() methodology. How would you
convert the 418978654 to a number that I could use as an index into a
fixed size array?

And what would the point of that be?

..NET already has collection classes that use the hash code directly (e.g.
Dictionary, Hashtable, etc.).

You can always just use the % operator to take an arbitrary number and map
it to some arbitrarily smaller range. But I don't see the point in this
case.

Pete

[Peter Duniho]

On Fri, 11 Jan 2008 16:15:26 -0800, j1mb0jay <no**@none.comw rote:

[...]
I just wanted to improve the insert method of the hash table because
after inserting all words I have to check a maximum 4 words rather than
1 to see if the word already exists in the hash table which is
2.5million words long.

There's no way for you to guarantee that you don't have to check multiple
words, even if you have as many entries in the hash table as you have data
elements. Collisions are always a possibility.

Besides that, you seem to be pursuing a pointless optimization. Even if
your hash table required you to check dozens of words, it would still be
ridiculously faster than a linear search of 2.5 million words. That's the
point of a hash table. You may be able to improve performance slightly by
implementing your own hash table and/or hash function, but IMHO once
you've got to the point of use _some_ kind of hash table, you're likely to
find the algorithm isn't going to get a lot faster by messing around with
the hash table. Your efforts will be better spent looking for other
low-hanging fruit elsewhere in whatever the algorithm you're doing is.
And if you can't get acceptable performance that way while still using a
hash table, you may wind up needing an algorithm that uses a data
structure that's even faster than the hash table.

Pete

[j1mb0jay]

Peter Duniho wrote:

On Fri, 11 Jan 2008 16:15:26 -0800, j1mb0jay <no**@none.comw rote:

>[...]
I just wanted to improve the insert method of the hash table because
after inserting all words I have to check a maximum 4 words rather
than 1 to see if the word already exists in the hash table which is
2.5million words long.

There's no way for you to guarantee that you don't have to check
multiple words, even if you have as many entries in the hash table as
you have data elements. Collisions are always a possibility.

Besides that, you seem to be pursuing a pointless optimization. Even if
your hash table required you to check dozens of words, it would still be
ridiculously faster than a linear search of 2.5 million words. That's
the point of a hash table. You may be able to improve performance
slightly by implementing your own hash table and/or hash function, but
IMHO once you've got to the point of use _some_ kind of hash table,
you're likely to find the algorithm isn't going to get a lot faster by
messing around with the hash table. Your efforts will be better spent
looking for other low-hanging fruit elsewhere in whatever the algorithm
you're doing is. And if you can't get acceptable performance that way
while still using a hash table, you may wind up needing an algorithm
that uses a data structure that's even faster than the hash table.

Pete

I am trying to do this for a data structure module for my degree,
writing our own data structures is important for best marks. More marks
are to be gained for good performance. Up to now from the data
structures i have studied this seems to be this best for this task. What
you do recommend for a better ?

j1mb0jay