How to control Transaction in task flow ?

Hi All,
In Each task flow under behavior tab we have transaction which contains following sub category
1-Always Begin New Transaction
2-Always use Existing Transaction
3-use Existing transaction if possible
Would any one give me real example which will show the purpose of this things
Thanks
Edited by: ADFORCLE on Jan 16, 2012 6:26 PM

http://one-size-doesnt-fit-all.blogspot.com/2011/05/jdev-11g-task-flows-adf-bc-always-use.html
http://one-size-doesnt-fit-all.blogspot.com/2011/05/jdev-11g-task-flows-adf-bc-one-root.html

Similar Messages

How to call an adf task flow in a commandNavigation tab

Hi,
I have created an adf task flow, I am able to invoke the same on click of a link.
Now I have a tab and some sub tabs under it. I want to open this adf task flow on click of one of these sub tabs.
For this I have added a property of type CommandNavigationItem in parallel with exising subtabs and on the actionListener of this property I am calling the method of the adf task flow.
But it is not working. When I click on the sub tab created by me I get this error on the logs :
ADF_FACES-60096:Server Exception during PPR, #1[[javax.servlet.ServletException
Please help/guide how to solve this issue.

Hi,
do you use the XML Menu Model? If so then the task flow must be referenced from a navigaqtion case in the unbounded task flow you turn into the menu model
Frank

How to create Popups in task flows

Hi,
Can any one help how can we create popup in taskflow. My requirement is i need to do create/edit operation on a table using task flows i have followed the "http://www.oracle.com/technology/obe/obe11jdev/11/btf/bounded_task_flow.html#t6" for create/edit operations in this it's navigating to new page, my case i need to invoke a popup instead of page.
can any one help me.

You need to create a separate tf for the edit and use that as a region in your search page.

How Worklist/BPM knows which task flow to invoke at runtime

Hi, I use JDev 11.1.1.5
In my BPM, I had one Human task activitiy state. Then I create a task flow through "ADF Task Flow Based on Human Task". I found that there is no any configuration value for taskFlowFileLocation property in .task source file. However, the process succefully invoke the correct task flow when I run it.
I am curious on how the worklist engine associate the task flow with the human task? If there are multiple task flows for the same human task, which file is the right place to find the configurations?
Thanks,

Hi Yannick,
I have a same scenario like yours. Have you implemented that? How to modify the preferences in backing bean? I change the input parameters of taskflow in backing bean. But when popup the edit dialog, the old value is still there. Thanks.
Regards,
Joe

How to control transaction isolation with tjdo + jboss

When running TJDO (a jdo-implementation) in conjunction with JBoss, tjdo's PersistanceManagerFactory is created by jboss as a service and can be retrieved using jndi when needed.
This tjdo is configured by a file tjdo-ds.xml in the deployment directory. Does anybody know if it is possible (and how) to control the transaction isolation level of tjdo in this case? The default is serializable, but experience indicates that using read_committed should be more useful if Oracle is used as database.
Best regards
Karl Brodowsky

this can be achieved by using
<config-property name="TransactionIsolation" type="int">2</config-property>

How to add many bounded task flow just on one only .jspx page?

Dear all
I will explain you what I do and What I want to do.
I created one jsf page template called "baseTemplate". It contains only one panelSpliter with horizontal layout . In the first face I created 10 commandLinks as a menu. the second facet contains the facet ref.
My task is now to create 10 task flows.
What i do is:
1- create bounded task flow name it "taskFlow1" and add all the views and page navigations that I want.
2- double click the views and create them as page fragment and design the layout.
3- create .jspx page name "task1.jpx" based on "baseTemplate".
4- dragging and dropping "taskFlow1" into "task1.jspx" as a region
5- repeating steps from 1 to 4 with "task2.jspx" till "task10.jspx"
in this way i created 10 "jspx" pages for each task flow
What i want to do is:
I want to create just one and only one ".jspx" page and drag all task flows on it. And when clicking on any command link from the menu , the corresponding task flow is invoke.
Is this a way to do this.
Thank you in advance.

of course, there is.simply in the unbounded Task flow drag your task flow as task flow call and specify the outcome, and make the action of your command link
the outcome you specify, of course you will navigate to the default activity of the bounded task flow also you could open your task flow as lightweight dialog
as frank do in his example

How to control transactions (WS-AT) between two web services in OSB?

Could someone tell me if the OSB can have 2 diferents web services controlled by an unique transaction?
For instance: in the OSB there are two wsdl, each one with its own Business Service and ProxyService respectively. In my vb.net aplication I started a transaction, called the first web service succefully and from the second web service I got an exception. I expected the first web service undo its work because the vb.net aplication excute a rollback command. Is this possible? How can I do? Any sample or tip will be helpfull.
If I was developing without OSB, I would rely on WS-AT as shown in http://www.codeproject.com/Articles/38793/6-Steps-to-Enable-Transactions-in-WCF for two web services dot net, or for one web service dot net and other java, http://www.ibm.com/developerworks/websphere/library/techarticles/0707_lo/0707_lo.html.
My question is basically: Is still possible to keep a single transaction between multiple web services when they are mediated by OSB? If so, is there a example showing how?

Hi,
When you create the proxy server, you have to check "Transaction required" box on Message Handling page...
http://docs.oracle.com/cd/E17904_01/doc.1111/e15867/proxy_services.htm#i1316487
Cheers,
Vlad

How tio navigate to parent task flow from Tab disclosure

Hi all ,
We have requirement is that ,
page contains two tabs
tab 1( Flow1 and Flow 2) user can navigate back and forth between flow 1 and flow 2.
tab2
from tab 1 user can navigate to flow 2 from flow 1 , if user navigates to flow 2 and modify some data and switch to tab 2 (click on tab 2) a warning message displayed to user stating that unsaved will be lost .
click in "Yes" user will be switched to tab 2.
from tab 2 if user click on tab1(switch back to tab1) , my requirement is that user should be navigated back to tab1 (flow1) not to flow 2.
Please suggest how can i implement this functionality.
thanks a lot for your help.

are you using the uncommitedDataWarning="on" property to check for dirtyness? if so... can you do that programatically by calling
ViewPortContext.isDataDirty();and based on this navigate back to the corresponding flows..
if i am not understanding ur usecase properly. can u pls explain a bit more..

Question about code corner Example#33: How-to open a Bounded Task Flow in a new Browser Tab

Hello All,
I am implementing example #33 in code corner series Oracle ADF Code Corner</title><meta name="Title" content="Oracle ADF Code Corner"><me…
Every thing is going fine, but there is a comment in the java code that I don't understand.
public void onLaunchForEdit(ActionEvent actionEvent) {
      //access the ADF binding layer and access the tree binding that
      //populates the table
        BindingContext bctx = BindingContext.getCurrent();
        BindingContainer bindings =
                               bctx.getCurrentBindingsEntry();
        //access the tree binding defined in the ADF binding layer
        JUCtrlHierBinding model =
               (JUCtrlHierBinding) bindings.get("DepartmentsView1");
        //get the current selected row key from the iterator binding
        //referenced from the table binding (tree binding). Of course,
        //I could have used the iterator name directly in the binding
        //lookup, but looking up the tree binding instead allows to
        //change the tree binding iterator dependency to a later point
        //in time without breaking this functionality. Its all about
        //"weak" dependencies that give you flexibility when coding ADF
        String rwKeyString =
               model.getDCIteratorBinding().getCurrentRowKeyString();
        launchWindow(rwKeyString);
What does this comment mean?
         //but looking up the tree binding instead allows to
        //change the tree binding iterator dependency to a later point
        //in time without breaking this functionality. Its all about
        //"weak" dependencies that give you flexibility when coding ADF
Is this a contrast to this line of code?
JUCtrlHierBinding model =
               (JUCtrlHierBinding) bindings.get("DepartmentsView1");
I mean in this line of code we hard coded the tree binding name from the page Def? why getting the tree binding name is a best practice while getting the iterator name is not?

What does this comment mean?
         //but looking up the tree binding instead allows to
        //change the tree binding iterator dependency to a later point
        //in time without breaking this functionality. Its all about
        //"weak" dependencies that give you flexibility when coding ADF
Is this a contrast to this line of code?
JUCtrlHierBinding model =
               (JUCtrlHierBinding) bindings.get("DepartmentsView1");
Yes, it's exactly this line of code the comment is referring to. If you look at the bindings of a page you see three parts: on the left he bindings, in the middle the executables and on the right side the data control. The statement
JUCtrlHierBinding model =
               (JUCtrlHierBinding) bindings.get("DepartmentsView1");
access the binding, the tree binding to get the data. if you use the iterator which named 'DepartmentsView1Iterator' you are accessing the executable, the iterator itself. The comment now tell you that if you access the data via the tree binding, that you can change the underlying iterator to point to different data. This you can do without the need to change the code in the been, whihc is good as you (or we all) tend to forget that we have code working on the iterator, so changing things in the bindings will break the application.
Timo

ADF task-flow and transactions

Hi All,
I have created a web application using ADF and JDeveloper 11.1.1.4. The ADF web application has two pages 'Search' and 'Edit'. The 'Search' and 'Edit' are page fragments. I have two bounded task flows 'search-flow.xml and 'edit-flow.xml'. The search and edit functionality has been created as dynamic regions by dragging the task-flows onto the JSF page.
The bounded search-flow.xml has page flow from search ----> edit(Parent Action component)
The bounded edit-flow.xml has page flow as *-->edit----> search(Parent Action component)
User comes from login page to the search page. On search page users can choose to search for records and edit individual searched records or create a new record using 'CreateInsert' button. Either way users end up on the edit page.
On the edit page, users see a form to fill or modify. They have the choice to navigate back to search page using 'Search' button or commit changed record using 'Save' button or create new record using 'Create' button
The application is functioning 90% as expected, only problem is the transaction management part if the user does not care to save each time the edit is made.
Say for instance, on search page user press ' CreateInsert' button which takes the user to edit page and then without saving the new record user presses 'Search' button which takes the user to search page and so repeated goes back and forth cycle.
In such cases the transaction never ends, The primary key for each new record is retrieved from the DB sequence only on persisting to database. A place holder like negative integer is used for the primary key temporarily till new record is persisted to database.
As per best practices, I am using the checkUncommittedDatabehavior which prompts the user when navigation is done away from the edit page i.e. without saving on pressing the 'Search' button on edit page. User can choose 'OK' or 'Cancel' which comes bundled with the ADF checkUncommittedDatabehavior property. Even after selecting OK the new record does not seems to be rollbacked (Issue 1).
eg:, if the user performs the cycle(i.e navigating between search and edit pages without commit/save) 'n' times, new entity records are created with primary keys -1, -2, -3, -4 ........ so forth. As I said earlier a proper sequence number for the primary key is assigned only if a record is persisted to the database.
Now on the 'n+1' cycle the user fills the records and clicks commit, it comes up with a error 'primary keys with nulls from previous uncommitted record (i.e; 1,2,3, .....n) creation.
I guess the transaction management is not happening correctly and not completing during these cycles (1, 2...n) Note: I am not using 'Task Return' components in my bounded ADF task flows, just commit operation 'Save' button in edit page and checkUncommittedDatabehavior.
Also, in search-flow.xml I have following configured under 'Behavior' tab Transaction ---> No Controller transaction and ' Share data controls with calling task flow checked
in edit-flow.xml, these configurations are in place under 'Behavior' tab Transaction ---> Always begin new transaction and ' Share data controls with calling task flow checked and Critical checked and 'task flow rentry' as reentry-outcome dependant.
Thanks
Edited by: user5108636 on Oct 26, 2011 5:16 PM
Edited by: user5108636 on Oct 26, 2011 5:24 PM

Hi Frank,
I checked the DB sequence is setup correctly. I have also modified the problem description above for better understanding.
2. When users press OK, navigate to a commit method activity (just drag and drop the commit operation).
-- When user presses OK, I need to rollback instead of commit to make users lose the unsaved work and controller should navigate to the search page. Here, the message panel with Cancel and OK is shown automatically with the <af:checkUncommittedDataBehavior/> property, I thought the framework will do the rollback for me. If I need to do rollback manually, how to do it, because I cannot see the JSF related source code for the <af:checkUncommittedDataBehavior/> dialog box. Please suggest what to do
3. If they press cancel, have the return activity rolling back to the save point taken when the user entered the bounded task flow (edit task flow)
-- If they press cancel, user stays on the same page, so I will leave future action to the user. This bit is working fine in my app.
Thanks
Edited by: user5108636 on Oct 26, 2011 5:23 PM

Task Flow Transaction Handling Inconsistent between JDeveloper R1 & R2.

We have a solution developed using Jdeveloper 11.1.2.1.0.
To describe the use case I will use the HR schema.
The user must be able to view Regions and navigate to Countries and Locations in a hierarchical way. At each level he must be allowed to view and maintain the detail of the current record at that level. The user must also be able to navigate directly to a level and perform a search to view the data and from there be allowed to to view and maintain the detail of the current record at that level.
So I have created the following design to achieve this and take into consideration that I need to reuse my task flows that maintain the data.
Business Services (Model Project)
1. HierarchyAM, containing Read Only VO’s of the various levels nested by association/view link.
2. Region + Countries + Locations
3. Search#Level#AM, A separate AM for each level, containing the Read Only VO per level.
4. Maintain#Level#AM, A Separate AM for each level, containing an updateable VO per level.
UI (View Controller Project)
I created the following BTF’s (To try and keep the post short as possible I will only provide detail for a single level to explain the issues)
*1. HiererchyBTF*
a.     Contains a fragment.
b.     Added RegionHierarchyBtf, CountryHierarchyBTF, LocationHierarchyBTF to the above fragment as dynamic regions.
c.     Task flow transaction <No Controller Transaction>, isolated.
*2. RegionHierarchyBTF. (The same BTF’s are created for each level in the hierarchy)*
a.     Contains a fragment.
b.     Exposed HierarchyAM<RegionRoVO>
c.     Task flow transaction <No Controller Transaction>, shared.
d.     Includes a call to the EditRegionBTF.
*3. SearchRegionBTF (The same BTF’s are created for each level in the hierarchy)*
a.     Contains a fragment.
b.     Exposed SearchRegionAM<RegionRoVO>
c.     Task flow transaction <No Controller Transaction>, isolated.
d.     Includes a call to the EditRegionBTF.
*4. EditRegionBTF*
a.     Contains a fragment.
b.     Exposed MaintainRegionAM<RegionVO>
c.     Task flow transaction <Use Existing Transaction if Possible>, isolated.
Runtime behaviour with above configuration.
1.     Running the HiererchyBTF and navigation the levels works in all versions.
2.     Calling an edit BTF allows me to edit the current record and successfully commit the data in all versions.
3.     Calling an edit BTF for the same record immediately thereafter, allows me to edit the record and commit the data successfully in all versions.
4.     Running any of search BTF’s and calling the edit BTF works in all versions.
Runtime behaviour and changing the EditRegionBTF task flow transaction to Shared.
1.     Running the HiererchyBTF and navigation the levels works in all versions.
2.     Calling an edit BTF allows me to edit the current record in the table and successfully commit the data in all versions.
3.     Calling an edit BTF for the same record immediately thereafter, allows me to edit the record and commit the data successfully in 11.1.1.4.0 & 11.1.1.6.0 only. If I perform the same scenario in 11.1.2.1.0 I receive the following exception.
a.     JBO-25003: Object MaintainLocationAM of type ApplicationModule is not found.
4.     Repeating the process, but after editing the first record, edit the second record in the table the error is not encountered. If, after successfully editing the second record and committing the data successfully, I again edit the first record I edited, I receive the same exception as in point 3.
5.     Running any of search BTF’s and calling the edit BTF works in all versions.
For the purpose of having a reusable edit task flow, I need to have the task flow transaction configuration as <Use Existing Transaction if Possible>, shared. This enables me to have the edit task flow as a parent as well as a child task flow and use the task flow functionality irrespective if it is controlling the transaction or not.
My questions are;
1. Is my design acceptable for the solution I am developing, I am more specifically questioning my separation of the AM’s to achieve loosely coupled task flows.
2. Is there a bug (I have done a search on MOS) that details this behaviour.
3. As, in my opinion, this is a common requirement, why is there inconsistency between the versions.
The demo for each version can be made available.
Regards.
Leon.
Edited by: Leon Dorfling on Feb 21, 2013 1:44 AM

Hi Frank,
+1. Its the first sting I stumbled on reading your post. Unless the read-only view object AMs are shared across users (application scope) I don't see why you want to have them in a separate AM. I don't see a benefit in a purely read-only view object that is not backed up by entities and thus would follow Oracle's recommendation to always use entities with view objects.+
The read-only vo's are based on eo's. I separated the AM as the search use cases are independent of one another. So my understanding, from a performance perspective, is to not have all the read-only vo's of the entire hierarchy of products in a single AM if I am only going to access one of them. If I do not understand correctly, can you please assist me in better understanding.
+2. For sure this behavior doesn't sound right and should be filed. You may want to check 11.2.3 in case it has been fixed in there already.+
As posted by Chris, I have through MOS found that it has been fixed in a later version.
Btw.: "Calling an edit BTF for the same record immediately thereafter, allows me to edit the record and commit the data successfully" how do you call this task flow dynamic region or task flow call activity ?
The BTF's are called using a backing bean.
Regards
Leon.

How to insert values in table using ADF Task flow? I am not able to bind CreateInsert on my first page fragement

Hi,
In Jdeveloper 12c I am trying to create a task flow where I have taken 3 pages to insert values in a table. On first page when I put data control of 2 columns as input with label, then on bindings tab I created an action as CreateInsert.
Now when I am trying to create invoke action in executables then this option is not available.
How can I create this task flow where values can be inserted in the table on the 3rd page fragment on commit button???
All 3 pages have data control input boxes to insert values in the table.

Hi,
Right click on the task flow page activity, click create page def and add there. If your activity is a page, then you can select the page and generate the pagedef as well.
Then you can add the executables.
Thanks.

How to create Discussions Task Flow Based on Oracle Content Server

Hi,
I need to create an ADF application consisting of the following aspects:
1) Upload, Download and Share documents (I'm using Documents Services provided by WebCenter Jdeveloper Extensions)
2) Create discussion threads and forums (I'm trying to use Discussion Services provided by WebCenter Jdeveloper Extensions)
3) User should be able to do an universal search i.e search both documents and discussions forums. (I'm using Search Services provided by WebCenter Jdeveloper Extensions)
I'm using "Oracle Content Server 10g for WebCenter" to manage my documents and discussions. While logging into GUI of OCS, I'm able to create and post discussion threads.
But, I'm unable to create discussions adf task flow based on Oracle Content Server.
The demos/tutorials/dev guides provided at http://webcenter.oracle.com have only described using WLS_Services managed server which is created and setup using WC Suite 11g.
But I couldn't find any information to develop these discussion services using Jdeveloper WebCenter Extensions and Oracle Content Server.
Please let me know, if anyone has tried this or if you have any idea of how to go about it.
Thanks in Advance,
Harikiran.
Edited by: harikiran on Jul 13, 2010 3:23 PM

Hi.
You have official documentation introducing ADF Task Flows here: http://docs.oracle.com/cd/E23943_01/web.1111/b31974/taskflows.htm
And you have next step by step tutorials to learn how to implement Task Flows:
http://yonaweb.be/webcenter_tutorial/part_6_building_taskflow_and_consuming_it_webcenter_portal (Yannick blog post).
http://docs.oracle.com/cd/E18941_01/tutorials/jdtut_11r2_57/jdtut_11r2_57.html (official JDev samples).
http://danielmerchanoracle.blogspot.com.es/2011/08/integrar-adf-task-flows-propios-en-una.html (My spanish blog with "how to add a Custom Task Flow into WebCenter Portal Application", 3 different and possible ways).,
Probably Yannick Blog post if you need to understand how to integrate Bounded Task Flows with WebCenter Portal's.
Regards.

How to create a task flow in webcenter?

Hi There,
I have a code fragment and that fragement is re-used across muliple pages in web center site. My requirment is to create a task flow based on this fragment.
I found some links but all are referring w.r.t WebCenter Spaces, which I am not intrested.
Please susggest me step by step guides or links so that I can referer and develop in webcenter.
Thanks in advance!!!
Regards,
Srinivas

Hi.
You have official documentation introducing ADF Task Flows here: http://docs.oracle.com/cd/E23943_01/web.1111/b31974/taskflows.htm
And you have next step by step tutorials to learn how to implement Task Flows:
http://yonaweb.be/webcenter_tutorial/part_6_building_taskflow_and_consuming_it_webcenter_portal (Yannick blog post).
http://docs.oracle.com/cd/E18941_01/tutorials/jdtut_11r2_57/jdtut_11r2_57.html (official JDev samples).
http://danielmerchanoracle.blogspot.com.es/2011/08/integrar-adf-task-flows-propios-en-una.html (My spanish blog with "how to add a Custom Task Flow into WebCenter Portal Application", 3 different and possible ways).,
Probably Yannick Blog post if you need to understand how to integrate Bounded Task Flows with WebCenter Portal's.
Regards.

Nested AM for Imported Bounded Task Flows - jdev 11.1.2.3 redhat 5.8

Hello:
Examples below all are Fusion ADF Web Apps.
subsystem1 = ss1 creates adflib1 with bounded task flow = tf1
subsystem2 = ss2 creates adflib2 with bounded task flow = tf2
Master system imports adflib1 and adflib2
In Master how can I create a app module that nests app modules from adflib1 and adflib2?
In Master, on a single page with splitter, I drop tf1 on one side and tf2 on the other side. I make a data change to tf1 and data change to tf2 and then commit. Changes are not commited on tf2. How can I make it to commit on both task flows 1 and 2?
I tried to create an app module in master and nest app module from ss1 and ss2 under that, but still changes are not committed on tf2.
Any ideas?
Thanks much.

Hi,
use shared data controls in your task flows so that a single transaction is created.
You want to read the following blog entries
http://bit.ly/MOVC1S
http://bit.ly/KOv5NY
http://bit.ly/KBCx0P
Frank

How to control Transaction in task flow ?

Similar Messages

Maybe you are looking for