Error conditions on tasks in a Workflow fork

Durrant,David William DURRAND2 at apci.com
Mon Jan 28 10:34:07 EST 2002


Hi Alon
 
As a beginner I probably didn't explain this very well.
We are getting the problems when a piece of data does not exist, and
therefore the task cannot even evaluate it to determine the outcomes.
 
So let's suppose we want to send an e-mail to an employee. Using the
Business object EMPLOYEET  we tell the task to send it to the Internet mail
address on this object. There are no outcomes on the send-email step so if
the data does not exist (because the Communications infotype has not been
completed for this employee), then it just stops, and holds up the entire
workflow, even thought the tasks are in seperate branches of a fork. If
there was some form of error trapping we could get it to skip the step and
hopefully let the other steps continue
 
The function we used was to extend the EMPLOYEET object with the attibutes
of the chief. It wasn't a seperate task in the workflow
 
Hope this is clearer
 
david
 
 
-----Original Message-----
From: Raskin, Alon (Soliance) [mailto:ARaskin at cps-satx.com]
Sent: 28 January 2002 14:51
To: SAP-WUG at MITVMA.MIT.EDU
Subject: Re: Error conditions on tasks in a Workflow fork
 
 
Hi David,
 
Every task in a workflow has an outcome. Most of the outcomes are the
exceptions that are defined on the method being called. For example, if the
method that gets the internet address returns an exception  then this will
show up in the Outcomes Tab of the Task display. You need to define a
corresponding 'path' on your workflow for that exception. To do that display
the OUTCOME tab on your task and enter some meaningful text. This will then
create a new 'path' in your workflow. If SAP Workflow encounters a task that
does not have a corresponding 'path' (I don't think that Path is the right
word but you get what I mean) then it will go to error mode. If there is a
defined path for that outcome then it will simply continue execution via
that path.
 
The other thing that I was wondering is why is the 'Get Email address' step
in parallel with the send mails task? Surely that should be the first thing
that the workflow does. Only after that is successfully achieved should the
workflow attempt to send the mail. If this fails (ie. No mail address is
determined) and no error occurs then the send mail step will error because
it did not have a recipient address. Just a thought....
 
Alon Raskin
Workflow Advisor - Soliance
(xtn. 3183)
 
 -----Original Message-----
From:   Durrant,David William [mailto:DURRAND2 at apci.com]
Sent:   January 28 2002 7:32
To:     SAP-WUG at MITVMA.MIT.EDU
Subject:        Error conditions on tasks in a Workflow fork
 
Dear workflow experts
 
We are using HR (PA and OM) and are new to workflow so we thought we would
start with something simple
We are using workflow to send e-mail notifications when a new employee is
hired into the company. We have the trigger events and starting conditions
figured out (so far). We built a fork into the workflow so that the emails
are sent in parallel in the belief (mistaken) that if one of the tasks
failed, then the other tasks would still execute. It seems we were wrong.
 
As an example
We are using a function to get the details of the person who is the chief of
the org unit, and then use the iinternet address stored on the
communications infotype on that person. This is one task out of seven in the
fork.
 
If the communications infotype does not exist then the whole workflow just
hangs.
 
Is there any way of setting an error trapping mechanism, to skip the task if
there is an error, and allow the other tasks to continue?
 
regards
 
Dave Durrant
Workflow Architect
Air Products PLC
 
+
 
+
 
 
+
 


More information about the SAP-WUG mailing list