how to generate unique Hash Code for string

[Ashish Khandelwal]

As MSDN is not giving us guarantee upon uniqueness of Hash Code, so
could any one suggest me that how to generate a unique Hash Code for
same string always, and generate different-2 Hash Code Different-2
string.

[Mannen]

"Ashish Khandelwal" <AK************ ***@gmail.comwr ote in message
news:e5******** *************** ***********@x29 g2000prg.google groups.com...

As MSDN is not giving us guarantee upon uniqueness of Hash Code, so
could any one suggest me that how to generate a unique Hash Code for
same string always, and generate different-2 Hash Code Different-2
string.

MSDN can't guarantee this, because it is not possible to give such a
guarantee.

Consider that the size of the hash is an int, and therefore 32 bits only (or
64 bits on a 64-bit system). This means that the contents of a string must
be "squeezed" into those bits, which in turn means a lot of data is lost.
For any given string bigger than the size of the hash, total uniqueness
cannot be guaranteed. I have never run into any problems with GetHashCode(),
but if you have, consider implementing your own MD5 (or whatever you feel
appropriate.)

[Ashish Khandelwal]

Thanks for reply.

but if you have, consider implementing your own MD5 (or whatever you feel
appropriate.)

Yes i am facing such problem, for exp. i am giving you "blair" and
"brainlessn ess" strings, these both strings are returning the same
Hash Code using GetHashCode() method, but as you said, use MD5, so can
you please give me some input related to MD5 like can i reply on the
output of this, and also as you were saying "or whatever you feel
appropriate" so here just same was my question that what should be the
way to get unique Hash Code, any idea...

[Jon Skeet [C# MVP]]

Ashish Khandelwal <AK************ ***@gmail.comwr ote:

As MSDN is not giving us guarantee upon uniqueness of Hash Code, so
could any one suggest me that how to generate a unique Hash Code for
same string always, and generate different-2 Hash Code Different-2
string.

Sure, so long as you've got an infinite range of numbers... My guess is
that you haven't though.

--
Jon Skeet - <sk***@pobox.co m>
http://www.pobox.com/~skeet Blog: http://www.msmvps.com/jon.skeet
World class .NET training in the UK: http://iterativetraining.co.uk

[Kerem Gümrükcü]

Hi Ashish,

use the Managed MD5 or SHA1 Class to make a unique
hash of your string. See this, it works with file streams,
but the technique is the same:

http://download.chip.eu/de/KHash-Tools-1.0_1317168.html

I wrote this, because i needed a very fast and console based
file hashing tool. There was none simple in www, so i wrote this.
It is open-source,...

Hashes generated with MD5 and or SHA1 are always unique!

Regards

Kerem
P.S: Where does your name come from...i am just curious,...

--
-----------------------
Beste Grüsse / Best regards / Votre bien devoue
Kerem Gümrükcü
Microsoft Live Space: http://kerem-g.spaces.live.com/
Latest Open-Source Projects: http://entwicklung.junetz.de
-----------------------
"This reply is provided as is, without warranty express or implied."

[=?ISO-8859-1?Q?Lasse_V=E5gs=E6ther_Karlsen?=]

Kerem Gümrükcü wrote:

Hi Ashish,

use the Managed MD5 or SHA1 Class to make a unique
hash of your string. See this, it works with file streams,
but the technique is the same:

http://download.chip.eu/de/KHash-Tools-1.0_1317168.html

I wrote this, because i needed a very fast and console based
file hashing tool. There was none simple in www, so i wrote this.
It is open-source,...

Hashes generated with MD5 and or SHA1 are always unique!

No, they're not

It's just very improbable that you can either find two files with
different content and the same hash, or manage to change a file and keep
its original hash.

But they're not unique.

The only guaranteed way to produce a unique hash for a stream of X bytes
is just to copy the entire stream.
--
Lasse Vågsæther Karlsen
mailto:la***@vk arlsen.no
http://presentationmode.blogspot.com/

[=?ISO-8859-1?Q?Lasse_V=E5gs=E6ther_Karlsen?=]

Lasse Vågsæther Karlsen wrote:

Kerem Gümrükcü wrote:

>Hashes generated with MD5 and or SHA1 are always unique!

No, they're not

It's just very improbable that you can either find two files with
different content and the same hash, or manage to change a file and keep
its original hash.

But they're not unique.

The only guaranteed way to produce a unique hash for a stream of X bytes
is just to copy the entire stream.

And of course then it wouldn't be named a "hash".

--
Lasse Vågsæther Karlsen
mailto:la***@vk arlsen.no
http://presentationmode.blogspot.com/

[Kerem Gümrükcü]

Hi Lasse,

>The only guaranteed way to produce a unique hash for a stream of X bytes is
just to copy the entire stream.

and thats what i asume here and why i wrote this!

Regards

Kerem

--
-----------------------
Beste Grüsse / Best regards / Votre bien devoue
Kerem Gümrükcü
Microsoft Live Space: http://kerem-g.spaces.live.com/
Latest Open-Source Projects: http://entwicklung.junetz.de
-----------------------
"This reply is provided as is, without warranty express or implied."

[Mannen]

"Kerem Gümrükcü" <ka*******@hotm ail.comwrote in message
news:Of******** ******@TK2MSFTN GP04.phx.gbl...

Hi Ashish,

use the Managed MD5 or SHA1 Class to make a unique
hash of your string. See this, it works with file streams,
but the technique is the same:

http://download.chip.eu/de/KHash-Tools-1.0_1317168.html

I wrote this, because i needed a very fast and console based
file hashing tool. There was none simple in www, so i wrote this.
It is open-source,...

Hashes generated with MD5 and or SHA1 are always unique!

No, they are not always unique, which is why they are called "one-way hash
algorithms". They are not reversible, because data is lost. The problem
still exists, but becomes less frequent (since MD5, for example, uses 128
bits instead of GetHashCode()'s 32 bits on a 32-bit system).

The reason for hash collisions is that n bits cannot be fit into x bits
where n x. Thus, when you hash a string of say 256 bits into 32 bits, 224
bits get lost because there is (obviously) no way to fit 256 bits into 32.
What you can do, however, is combine the 256 bits into the 32 bits in a
certain way (ie, the algorithm) to make it more likely that a unique
bitpattern is produced. This is what the hashing algorithms are all about.

http://en.wikipedia.org/wiki/MD5

>
Regards

Kerem
P.S: Where does your name come from...i am just curious,...

--
-----------------------
Beste Grüsse / Best regards / Votre bien devoue
Kerem Gümrükcü
Microsoft Live Space: http://kerem-g.spaces.live.com/
Latest Open-Source Projects: http://entwicklung.junetz.de
-----------------------
"This reply is provided as is, without warranty express or implied."

[Kerem Gümrükcü]

Hi Mannen,

>What you can do, however, is combine the 256 bits into the 32 bits in a
certain way (ie, the algorithm) to make it more likely that a unique
bitpattern is produced. This is what the hashing algorithms are all about.

Sure, then you have to implement your own hashing class that works with
MD5 as its base. I am familliar with MD5 and SHA1, i know its "pitfalls", ...

Regards

Kerem

--
-----------------------
Beste Grüsse / Best regards / Votre bien devoue
Kerem Gümrükcü
Microsoft Live Space: http://kerem-g.spaces.live.com/
Latest Open-Source Projects: http://entwicklung.junetz.de
-----------------------
"This reply is provided as is, without warranty express or implied."