Performance of OO vs BOR as used in WF

[Sample, Rick]

Hey Mike, 

We will be moving to ECC6 on our business side soon. With our HR move, 
we barely had time to stomp out conversion bugs much less roll 
the sleeves up and explore the new ABAP objects. 

I have an old ABAP Objects book collecting dust. But, when we get there 
I would like to have my resources available.

What SAP OO books are on your self? 

Rick Sample
SAP Business Workflow Developer
Graybar, Inc.


-----Original Message-----
From: sap-wug-bounces at mit.edu [mailto:sap-wug-bounces at mit.edu] On Behalf
Of Mike Pokraka
Sent: Friday, March 21, 2008 4:22 AM
To: SAP Workflow Users' Group
Subject: Re: Performance of OO vs BOR as used in WF

Hi Alon,

Short answer:
OO has the potential to be a lot more efficient, but it's your
responsibility to realize the potential.

Long answer:
The beauty of classes is that the answer to this question is in the
hands
of the developer. In a pure BOR vs OO context, there is very little
difference in the internal WF code. Everything is accessed via Local
Persistent Object References (LPOR).

The difference is that BOR has a fixed overhead (SWC_CREATE_OBJECT) that
you cannot change or control. With OO, the WF system simply calls
BI_PERSISTENT~FIND_BY_LPOR. What you put in there is up to you. You can
have a SELECT for each instantiation or you can implement a more
efficient
buffered instance management.
You can even get really clever and pre-select data you need for other
instances so that one SELECT serves multiple objects.

Whilst the BOR instance management is reasonably efficient, it is still
an
old cludge to make a procedural language object oriented, so will
defeinitely have some performance drawbacks that are not necessary or at
the very least cam be avoided in some cases. e.g. it will always do a
"SELECT *" for all DB attributes and the key table, whereas you can
implement a "SELECT <fields you need>" in your OO class.

Cheers,
Mike


On Wed, March 19, 2008 4:08 pm, Alon Raskin wrote:
> I am curious if anyone has run any tests to see whether performance is
> different when using OO container elements (class instances) as
opposed to
> BOR (object type instances). My gut feel is that OO would be faster
due to
> the reduced overhead of not using those BOR macros but I am curious if
> anyone has actually done the comparison.
>
> Alon