If its stats then analytics will likely be the best way to go. Failing that
implementing in JAVA and executing the java in the db would seem a
reasonable approach.
--
Niall Litchfield
Oracle DBA
http://www.niall.litchfield.dial.pipex.com
"Christian Tischler" <christian_at_myunix.net> wrote in message
news:dbfg20$n6l$1_at_online.de...
> Rob Wolfe wrote:
>> I would strongly suggest that you take a long hard look at the
>> analytical functions available in oracle. There is almost no chance
>> that you will have to completely re-invent the wheel.
>>
>> I have to agree with Jim that I have never had to write C/Java/C++
>> (whatever) to increase the speed of an oracle query and quite honestly
>> cannot imagine the circumstance where i would have to for THAT
>> particular reason.
>>
>> I would also suggest that you first get a good grounding in how
>> Oracle's flavour of SQL can help you before learning PL/SQL. I often
>> find that folks do not think of things in the kind of set-oriented ways
>> that make SQL powerful and instead do as you originally suggested and
>> use a RDBMS as simply a storage mechanism. If that is all you are using
>> it for you are spending a large amount of money for very little return
>>
>> Rob
>>
> That is the reason why I asked questions here, I wanted some hints in the
> right direction. Until now I used free DBs so no monex was spent so far.
> Oracle was chosen over MS SQL as these two are the only options supportet
> by be "DB feeding" complex.
> We need to implement rather complex statistical weighting functions and
> stuff like that. I will hopefully find something in Oracle that will make
> this job easier, as we certainly do not plan to reinvent the wheel.
>
> thanks again
>
> Christian Tischler
Received on Mon Jul 18 2005 - 15:21:50 CDT
