I have a curious problem with an ASP.NET user-control which is dynamically created in VB.NET. Sorry that this is a long post, but I think it’s better for me to explain carefully what I’m doing and what the problem is.
Essentially, the user-control is an HTML table used to display data from a back-end SQL Server DB. It uses a stored procedure to return a DataTable, from which it builds html rows and columns with the values stored in the InnerHtml property of the relevant table cell. This is just a passive display of the data and works perfectly.
However, the user-control also has the facility to insert a new record, triggered by an ASP button. This causes the following sequence of events;
Post-back
The user-control_Init event rebuilds all the existing rows in the table and inserts their values as described above.
The page_Init event.
The page_Load event.
The user-control_Load event.
The button_Click event then builds a new row with the same structure as the main table. But this time it inserts into each table cell a relevant ASP.NET control, such as a textbox or checkbox, holding “null” values. Each control has its ID set explicitly to a unique value. Finally, an “insert flag” is set to true.
The user then enters the data into these controls and clicks an Insert-Save button. This causes another post-back, which follows the procedure described above, except that the “insert flag” tells the user-control_Init event (2) to rebuild the insert row. Then, before the page_Load event (4) the values previously entered by the user are reinstated in the ASP.NET controls from ViewState. Finally, the Insert-Save_Click event calls a stored procedure to save these data to the DB.
All this work perfectly when there are existing rows in the html table. But – and this is the strange behaviour – it fails when there are no existing rows. The first post-back completes successfully although, obviously, event 2 skips over the rebuilding of the existing rows because there are none. However, the insert row is rebuilt correctly and the user can then enter data into the ASP.NET controls. During the second post-back (Insert-Save), the first three events described above work perfectly. But, the moment the code hits the page_Load event (4), all the ASP.NET controls disappear. The html row and cells are still present, including the literal control (index 0) in each cell. No error is thrown at this stage, but when the Insert-Save_Click event occurs and the data should be read from the ASP.NET controls – they are not available and an "index out-of-bounds" error is thrown, because the ASP.NET controls should each have an index of 1 within their cells.
I’ve spent hours (days!) debugging this and I cannot see what is causing these controls to just evaporate! Any clever programmers out there got any ideas please?
Related
vb.net, .NET 4.0, sql server 2008 backend.
I have a webpage which contains 3 formviews, each within their own Panel. As the user traverses the page, depending on which step they're on, a certain formview will be visible. After the process is complete, the user will click the Submit button and The button onclick event calls each FormView.Insert statement. Each Formview inserts into a separate table. The issue is that if the last FormView.Insert throws an exception, the first 2 Inserts have already been executed successfully.
My question is, and I hope this isn't too vague but...What is my best option at this stage?
1. Create a single Stored Procedure to handle each Insert in order to handle if one of the Inserts throw an error?
Thank you.
I'm having an issue binding the value of a page item to a declared variable in an anonymous PL/SQL block process.
The problem is that the page item (:P4550_REQUESTOR) is not populated with a value until a conditional is met. It appears that the PL/SQL block process is binding the variable to an empty value as soon as the page is loaded, despite the fact that the process does not fire until a specific button has been clicked.
Here is my code:
DECLARE
v_email_to app_user.email%type;
v_requestor VARCHAR2(15);
BEGIN
v_requestor := :P4550_REQUESTOR;
BEGIN
SELECT email INTO v_email_to
FROM app_user
WHERE userid = v_requestor;
END;
SEND_APEX_MAIL (
v_email_to,
'Your vacancy request has been rejected.'
|| chr(10)
|| 'Emailed to: ' || v_email_to
|| chr(10)
|| 'Requestor: ' || v_requestor,
'Vacancy Request Rejected'
);
END;
Does anyone have any thoughts on this?
The block works just fine if I hard code a value to v_requestor. If I try to get the value of P4550_REQUESTOR after the page has loaded, it is empty. After clicking the edit button, P4550_REQUESTOR is populated.
** **MORE DETAIL** **
P4550_REQUESTOR is a page item that resides within the Vacancy Request region which is only displayed when a conditional is met. Specifically, the conditional is an edit button associated with a table row that is created on page load. Clicking the edit button causes the details region to display, and the associated page items to be populated.
The page item values in the Vacancy Request region are populated via an Automated Row Fetch which fires After Header.
P4550_REQUESTOR has a Source Type of DB Column.
The process that fires the code above is set to fire On Submit - After Computations and Validations
If I log the value of P4550_REQUESTOR when the page loads, it shows null. If I log the value after clicking the edit button, I get the expected string value.
Process Flow Control in Oracle APEX
(This is actually useful to think about in other programming disciplines and environments.)
Problem Defined
The problem is that the page item (:P4550_REQUESTOR) is not populated with a value until a conditional is met. It appears that the PL/SQL block process is binding the variable to an empty value as soon as the page is loaded, despite the fact that the process does not fire until a specific button has been clicked.
The problem statement reworded in Apex terminology and presented in the form of an actual question:
There is a REPORT REGION on the page which contains the result of a direct reference to a data table/view. This report is managed by an Apex process called "Automated Fetch" and is initiated automatically by the loading of the page headers.
There is a FORM ITEM on a page which which is populated conditionally by a BUTTON ITEM selection made by the user. The BUTTON ITEM is part of the report results.
There are multiple button items. Each is associated with a value for each report record.
If the user does not select the BUTTON ITEM from the REPORT REGION, the FORM ITEM remains unassigned and contains a "null" value.
There is a defined PL/SQL block of code which is set to execute when a SUBMIT BUTTON item is pressed (also on the same page). Why does my code block (defined page process) run with a null value when it is triggered without first pressing a BUTTON ITEM from the REPORT REGION first?
Event Driven Program Design for Procedural Programmers
The answer is not obvious if you think under the paradigm of a procedural language. Without diving into a lecture on the topic, here's a visual layout of the problem space of the OP that I cooked up to illustrate how the problem can be made more obvious:
This is my Apex page design in implementation. It's generic enough to use as a template for other Apex designs. There are no flow arrows on this diagram because it's a stateful system. One thing causes another thing to happen and so on... but not always and not all at the same time.
Use Cases for Apex UI Page Designs
Try walking through a few use cases to understand how the elements broken down in the diagram operate together. Each user may take any number of click combinations and interactions, but there is a commonality:
They all enter the same initialized conditions on page load.
They all leave the page by: navigating elsewhere or through the SUBMIT button event.
Use Case #1
User chooses {MyPage:SQLReport:ThisButton} from one of the records in {MyPage:SQLReport}
According to {MyPage:SQLReport:ThisButton} #3, the value associated between the report record and the button item is passed to: {MyPage:HTML-Region:ThisItem}
The form item state has been updated and changed from the initial null value.
User selects {MyPage:HTML-Region:ThisSubmit} button to inform the system to continue on.
The submit button executes the defined PL/SQL procedure block: {MyPage:RunCodeBlock}
Use Case #2
User enters page and reviews results displayed in the {MyPage:SQLReport} region.
User decides no additional input is necessary and then selects the {MyPage:HTML-Region:ThisSubmit} button to inform the system to continue on.
(a note: the state of form item {MyPage:HTML-Region:ThisItem} has not been changed from the initial null value at this point... after the submit button has been selected)
The submit button executes the defined PL/SQL procedure block: {MyPage:RunCodeBlock}
Use Case #3
User chooses {MyPage:SQLReport:ThisButton} from one of the records in {MyPage:SQLReport}
According to {MyPage:SQLReport:ThisButton} #3, the value associated between the report record and the button item is passed to: {MyPage:HTML-Region:ThisItem}
The form item state has been updated and changed from the initial null value.
User chooses {MyPage:SQLReport:ThisButton} from a different selection from one of the records in {MyPage:SQLReport}.
According to {MyPage:SQLReport:ThisButton} #3, the value associated between the report record and the button item is passed to: {MyPage:HTML-Region:ThisItem}
The form item state has been updated and changed from the initial value stored in step (2).
User selects {MyPage:HTML-Region:ThisSubmit} button to inform the system to continue on.
The submit button executes the defined PL/SQL procedure block: {MyPage:RunCodeBlock}
The difference between each case should illustrate why the dependent value (ThisItem, or more specifically, page item P4550_REQUESTOR) is null in one use case vs. the other.
Building a Physical Implementation (An Apex Page)
The table I used is called STAR_EMPS. It is similar to the EMP table but has only three columns: ename, deptno and salary. Although it is not super important, this is the data set I used to populate STAR_EMPS:
I used a simple two-column table named STAR_EMPS_LOG for capturing the output of a successfully executed procedure call. You could accomplish the same with just one column, but I wanted a sequential id for tracking the order each event was recorded- for running multiple test cases. The procedure is one of several defined processes kept on this page:
contained in: {MyPage:RunCodeBlock} is below:
DECLARE
-- output from this procedure will be recorded in the star_emps_log
-- table. {MyPage:RunCodeBlock}
mycelebrity star_emps.ename%TYPE:= :P17_CELEBRITY_NAME;
mylogmessage star_emps_log.log_message%TYPE;
BEGIN
-- Conditional; changes message based on the value set for the
-- page item.
if mycelebrity is null then
mylogmessage:= 'No button was pressed on the previous page.';
else
mylogmessage:= 'The user selected: ' || mycelebrity ||
' from the report list.';
end if;
-- populate value from the page item.
INSERT INTO star_emps_log (log_message)
VALUES (mylogmessage);
commit;
END;
This is how the page layout was set up:
As in your example, I made a {MyPage:SQLReport} region with its supporting elements. The SQL Report represents a query directed at the source data table.
{MyPage:Form} has been renamed to {MyPage:HTML-Region}.
{MyPage:SQLReport} is defined by a SQL query, there is also a mock column to use as a place holder for placement of the "edit" buttons.
{MyPage:SQLReport:ThisButton} The button specifications are detailed through this:
The TWO Page processes: PROCESS and BRANCH need to be linked with the same settings referencing a BUTTON triggering Item.
User Interface Test Cases
Run through the three suggested scenarios to get started. Verify that the system is interpreting the requests correctly. This is what the page layout looks like:
The two processes on the system have a definition that wasn't mentioned in previous discussions may solve our original problem at hand:
Some Parting Thoughts
It is a good thing this turns out to be a trivial case once broken down. The diagramming method described here should scale to other Apex applications of varying complexity. There is considerable utility in stepping away from the code, locking down on terminology and trying to describe systems and processes without actual code. Please be sure to share any stories if this approach helps with your own Oracle Apex design challenges.
Onward!
The original, verbose answer seems to way overcomplicate the issue. The session state concepts manual covers this behaviour more succinctly.
Should P4550_REQUESTOR be a normal item created from a wizard, using :P4550_REQUESTOR will return a value in processes running post submit because the submit processes moves values in browser to session state.
If P4550_REQUESTOR is rendered conditionally, then it will always be null and I'm not sure what would happen if you tried to set it - probably depends how.
On a similar note, if you used &P4550_REQUESTOR. to parameterise the process, you would face the behaviour originally described (and made the code less secure)
Suppose I have a custom control like a Table with a varying number of rows in it. What would be the advantages of using databinding (like in a GridView) if I went that route? What would be the advantages if I used the control state to keep track of the number of rows?
Edit: Example.
Each TableRow in the Table object represents an employee. Each TableCell contains a TextBox input control. The user can add another blank employee (row) with a button under the Table. Information entered must be maintained through postbacks.
Using the states method, the number of rows can be saved into the control state. Since control state is loaded before view state, the number can be restored before the view state is restored (providing objects to take the view state and post values).
Binding with a GridView? I don't know much about how this works. From what I've seen, you can XML-serialize the data and save it to the Value property of a Hidden control.
Within AX 2009, I have, through compare and compile, added two new controls within a datagrid on a form, a Real edit and a combobox. I have compiled with no issues. The Allow Edit property is set to Yes on both controls.
However, on the form, if I edit one row, whether typing a new number with Real edit or combobox, and don't hit Save but hit the Down Arrow key, the data I typed on the previous records is duplicate in the next record and so on until I release the Down Arrow key, rather than just setting the focus on a new record.
The table where these fields were created doesn't exhibit this behavior. The focus simply moves to the next record and what was typed will not carry over to the next record. Only the form does this...
Has anyone seen this behavior before with AX forms?
You may have omitted to specify the data source on the grid itself?
Or if the new controls are based on Edit methods on the data source, have you got the data source parameter in the method signature?
see http://msdn.microsoft.com/en-us/library/aa637541(AX.10).aspx
I'm looking for the easiest way to bind data from a SqlDataSource to textboxes dropped in Visual Studio 2008.
For example, I have 4 textboxes currently that have Address, City, State, Zip. I also have a SqlDataSource on the page fetching the ID of the record and selecting those 4 fields based on ID.
How am I able to quickly bind each box to those particular fields selected? I would think this would be really straight forward - but seems it's not. Seems like the answer is funneled towards having to create a GridView or some type of control.
Be gentle...I'm a nub :)
In general you are correct, if you want to use databinding you'll need to use an appropriate control. For this example I'd suggest using a FormView - it is designed to display the results from a single database record and uses templates, meaning you'll have complete control over the output. This article is probably a good place to start: FormView Control: Step by Step.
To read the values bound to the FormView in the code-behind class you would need to create an event handler for the FormView's DataBound event. In that event handler you would reference the controls programmatically via FindControl, like so:
Dim myLabel As Label = CType(FormViewID.FindControl("id"), Label)
Here, id would be the ID of the Label whose value you were interested in. Once you have a reference to the Label you can get its value using myLabel.Text.