Re: Will ROWID change during DB Runing??

In article <crdbkt$qg3_at_netnews.proxy.lucent.com>, Grant Zhou says...
>
>Hans,
>Thanks very much!
>I am reading the books you mentioned.
>
>But I only want to get one answer:
>Will the ROWID change for a record if we only update the data? T/F
>:?
>Best Regards,
>Grant

if you "enable row movement" -- yes, it can change due to a simple update (believe someone posted that with an update to a partition key)

with IOTs the "rowid" is a fake one, updating the primary key will change the "rowid" of an IOTs

in 10g - lots more stuff let you use enable row movement to do things -- so yes, the ability for a rowid to change over time will become more and more frequent

>
>"HansF" <News.Hans_at_telus.net> wrote in message
>news:pan.2005.01.04.02.56.39.959256_at_telus.net...
>> On Tue, 04 Jan 2005 10:41:38 +0800, Grant Zhou wrote:
>>
>> > The program caused 3 user of 80K user charge error. I have doubts about
>the
>> > rowid usage.
>>
>> ROWID is very efficient if used correctly. However, it is not necessarily
>> the correct technique to access a row, especially if there is an
>> intervening commit.
>>
>> I agree that you are possibly seeing an example of incorrect usage.
>>
>> WHen you are ready to go to a supported version of Oracle (either 9i or
>> 10g), add "Effective Oracle By Design" to your list of reading material.
>> (Actually, a lot of the material is relevant for you right now, but I'm
>> encouraging you to move up to a supported version. <g>)
>>
>> /Hans
>>
>>
>
>

-- 
Thomas Kyte
Oracle Public Sector
http://asktom.oracle.com/
opinions are my own and may not reflect those of Oracle Corporation