Re: hash value - function

From: Brian Peasland <dba_at_nospam.peasland.net>
Date: Tue, 14 Aug 2007 13:36:19 -0500
Message-ID: <46c1ea0c$0$1099$88260bb3@free.teranews.com>

fitzjarrell_at_cox.net wrote:
> On Aug 14, 12:19 pm, ciapecki <ciape..._at_gmail.com> wrote:

>> On 14 Aug., 18:21, "fitzjarr..._at_cox.net" <fitzjarr..._at_cox.net> wrote:
>>
>>
>>
>>
>>
>>> Comments embedded.
>>> On Aug 14, 10:54 am, ciapecki <ciape..._at_gmail.com> wrote:
>>>> Hi,
>>>> Is there any function (10g) that would return a kind of hash value,
>>> Which release of 10g?
>>>> I was thinking of getting one value for the whole table.
>>> I'm wondering why you'd think to do that.
>>>> My idea was to concatenate all the columns (5 columns there not very
>>>> big around 10chars each) get a numerical hash value.
>>>> Then group by primary_key, and get a sum or avg, or whatever.
>>> What possible benefit could that provide?
>>>> If the hash value would be big enough I could use it for comparing if
>>>> two tables are the same of the content or not.
>>> How?  Methinks this be folly of a grand order.
>>>> the question is, if there is such a function, I could not find it that
>>>> takes varchar2 and returns a value.
>>> DBMS_CRYPTO.HASH_MD4
>>> DBMS_CRYPTO.HASH_MD5
>>> DBMS_CRYPTO.HASH_SH1
>>> are available,  but they won't hash an entire table.
>>>> thanks
>>>> chris
>>> I would be seriously re-thinking this idea; what business case does it
>>> support?  I have a strong feeling this is a path which is rapidly
>>> going nowhere.
>>> David Fitzjarrell
>> The idea is the same as zipping and unzipping files, the software
>> creates a kind of hash value for the file, and after unpacking it
>> checks if the new counted hash value is the same that was originally,
>> and then it means we got the same file.
>>
>> I wanted to use this procedure to check periodically if a table
>> changed.
>> I would save the original hash value, and run periodically checks on
>> the table if the hash value is different,
>> if it is the case it means the content of the table changed.
>>
>> chris- Hide quoted text -
>>
>> - Show quoted text -

>
> Why go about this convoluted act? The USER_TAB_MODIFICATIONS view
> provides such information; a simple process of monitoring that view
> and reporting any changes should suffice. You have still not answered
> the question of what business case this satisfies; my guess is none.
> To 'hash' an entire table simply to see if one column in one row has
> changed is exerting far more effort than the task is worth, in my
> opinion.
>
> I would be investigating the tools already in place which Oracle has
> provided, rather than scheming to devise new and devilish mechanisms
> which are prone to failure and impact the database and user community
> in a negative manner for absolutely no gain whatsoever.
>
>
> David Fitzjarrell
>

I can't imagine hashing a 100MB table....ouch!

Cheers,
Brian

-- 
===================================================================

Brian Peasland
dba_at_nospam.peasland.net
http://www.peasland.net

Remove the "nospam." from the email address to email me.


"I can give it to you cheap, quick, and good.
Now pick two out of the three" - Unknown

-- 
Posted via a free Usenet account from http://www.teranews.com

Received on Tue Aug 14 2007 - 13:36:19 CDT