GreyBeard wrote:
> On Mon, 21 Feb 2005 21:23:23 +0100, Frank van Bortel wrote:
>
>
>>>Please correct my interpretation of your statement:
>>>
>>>"Selecting from inappropriate views that do unnecessary work [relative
>>>to what is needed by the select] can lead to bad execution plans.
>>>
>>>Therefore selecting from views is bad."
>>>
>>
>>No - I don't go that far; I'd like to change that to:
>>therefore selecting from nested views should only be done
>>with caution by knowledgeable developers.
>>
>>>In some circumstances (such as publishing views to ad-hoc users) I can
>>>concur that data sources can be used inappropriately.
>>
>>Meet Mr. Murphy! Mr. Murphy is an optimist.
>
>
> I rate this on the same level as indexing - my feeling is that the
> continued existance of every [performance related] index should be
> rejustified every 6 months.
>
> Back to this one - would it be fair to provide the following ROT to the
> OP's question:
>
> "The technical issues that caused performance concerns related to nested
> views in [very old] versions of Oracle are, for the largest part, no
> longer relevant.
>
> However, inappropriate use of views can have side effects such as
> unacceptable performance and resource issues. These side effects, such as
> unnecessary aggregation within views, could become significant in ad-hoc
> environments.
>
> Suitable documentation for views as well as proper analysis, including
> explain plans, should be considered for all 'permanent' SQL statements."
>
> FGB
I like that - it may just be a little long for a ROT.
The part that sticks will probably be just the first
sentence, which is OK by me for a ROT.
--
Regards,
Frank van Bortel
Received on Tue Feb 22 2005 - 03:19:52 CST
