Session state and PDF building

Trying to get a application report to use a few variables in a report being built from APEX. I have gone through and built the application report and SQL, I clicked on 3 items that are in session state, to push them through to the XML file being produced.
However, when I open that xml file I find the following:
<P167_END_DTE></P167_END_DTE>
<P167_FACILITES></P167_FACILITES>
<P167_START_DTE></P167_START_DTE>
The 3 session items are NOT returning values to the XML file produced. Is there a possibility I have missed something?
Using Apex 3.0.... I have selected the Advance XML structure (to include session information)
Please advise..
Thank you,
Tony Miller
UTMB/EHN

Tony & Costantino:
When downloading the XML you don't have access to the session state of your application. Only when running you application will the underlying XML structure actually be populated with your session state values. So it's expected that your session state item XML tags do not have any values at design time / when creating the report query. They're just there so that you can pick them up with the BI Publisher Word-Plug-In and place them in your RTF template. For that you don't need values.
Regards,
Marc

Similar Messages

Session state and browser cache - Back button problem

Hi all,
I have a problem (and unless I'm missing something I think we all do) with session state and use of the browser's Back button. I really hope I'm just being dumb...
Background scenario:
Page P has a sidebar list allowing the user to select what content is displayed (e.g. 'stuff relating to X, Y or Z' where X, Y and Z are rows in, say, a table of projects). When a list entry is clicked, we branch to page P with the value of the list item placed in an application-level item (call it G_PROJECT). Reports on page P use G_PROJECT in their WHERE clauses.
So, click list entry X and G_PROJECT is set to X and page P shows reports for project X.
Page P also has a set of buttons which branch to various edit pages which allow attributes of page P's current project to be updated. These pages similarly use G_PROJECT in their WHERE clauses.
Problem scenario:
1. The user goes to page P and picks project X off the list. Project X's stuff is displayed (G_PROJECT = X).
2. The user then picks project Y off the list. Project Y's stuff is displayed (G_PROJECT = Y).
3. The user then clicks the browser's Back button. The page is served from browser cache, so project X's stuff is displayed, but G_PROJECT still = Y.
4. The user clicks an 'Edit' button; we submit, and branch to an edit page which displays (and will edit) data for project Y because G_PROJECT still = Y.
This is SERIOUSLY BAD NEWS - apart from being confusing, the user's edit permissions on projects X and Y may differ, and so the user may be able to perform 'illegal' updates.
I've read what I can on this forum and the rest of the web looking for ways to a) inhibit browsers' 'Back' functions and/or b) prevent pages being cached by the browser, but none of them have worked for me.
Short of waiting for browser manufacturers to recognise that the web is now full of applications as well as static pages, and enable robust programmatic control of cache behaviour, does anybody know how the problem can be avoided - or at least detected?
Thanks,
jd
Failed attempts to date:
<meta http-equiv="cache-control" content="no-cache">
<meta http-equiv="cache-control" content="no-store">
<meta http-equiv="cache-control" content="private">
<meta http-equiv="cache-control" content="max-age=0, must-revalidate">
<meta http-equiv="expires" content="Wed, 09 Aug 2000 01:01:01 GMT">
<meta http-equiv="pragma" content="no-cache">
Disallowing duplicate submission (page attribute).
window.history.go(1);

Thanks Scott,
I may be being dumb here but I don't see how that would help...
P250_PROJECT and G_PROJECT are currently kept in sync by app logic. Whichever is used to drive, if the page is rendered from cache then the app logic is not executed, so the rendered page contents are not those keyed by P250_PROJECT, as illustrated in steps 1-4 of the problem above.
The user sees X, the session items say Y. The engine doesn't know what the user is seeing.
when page P is POSTed, its hidden item P250_PROJECT should always be used to derive the application item G_PROJECT. Then whether the page was pulled from cache or rendered anew via a click from the sidebar link, the project ID is determined by the contents of that page.
As I said above I tried this, with the 'Edit' branch set to:
Set these items: G_PROJECT
With these values: &P250_PROJECT.
but it makes no difference. The project ID is not determined by the rendered page contents - the engine gets the value of P250_PROJECT from session state.
I can code the 'Edit' pages such that they check permissions and if necessary redirect back to p250 (conditional before-header branch), but that's a clunky cure rather than the prevention I was hoping for.
Please tell me if my understanding is incorrect.
jd

Session state and item values

Hi,
Is it any possibility to bring the session state and an item value to a table?
Thank you

Hi,
For session id use APP_SESSION substitution string
http://download.oracle.com/docs/cd/E23903_01/doc/doc.41/e21674/concept_sub.htm#sthref160
Regards,
Jari
http://dbswh.webhop.net/dbswh/f?p=BLOG:HOME:0

Controlling Session State and Submit Processing

I am trying to gain a fundamental understanding of how to control submit processing and session state, etc.
For example, I have a popup lov that lists names and returns id for the parent record. When the user make the selection using this pop lov, I then have an id that I want to use to lookup the default value for another field.
How do you submit a page for re-rendering (i.e, you have condition expressions referencing values that the use has set) without triggering the auto dml and page branch?
For example, I have a form for specifying tns names with 2 buttons, 'Advanced Config' and 'Basic Config'. The idea is to hide/show items corresponding to basic or adv. config options, etc.
My first approach was to have the buttons do a redirect to the same page, setting P5_USE_ADV_CONFIG_YN to Y or N without clearing the cache. When the page refreshes, the conditional rending has occured as desired (i.e., I set the condition on my form items to reference P5_USE_ADV_CONFIG = Y or N, etc.), but user's data values have been lost since the record was reloaded from the database.
So..I tried putting a condition on the fetch procedure to prevent the record from being re-read. However, this left all the fields empty which indicates that the item's session state has been completed cleared! Huh? I didn't want ask for or want this behavior. Hmmm.
So, I start thinking that the persisted session wasn't updated since I didn't do a submit and so when the page was requested, it behaved as it should, but not what I had in mind. THUS...I believe I need to submit the page and prevent the auto-dml and page branching afterwards....but how? I feel that I am missing something very fundamental here and that the answer may also help me with other related questions:
* How does the confirm delete examples work? How is the processing stopped when the user presses cancel?
* How can you do validations and stop dml processing without losing the user values?
* How can you display and error page and then return to where the user was with the values still present?
* How can I do a secondary sql value lookup based on the users selection in a popup lov? (User selects a Database/DbID - then I want to lookup host using DbId)
I haven't been successful understanding this from the manuals, but someone can point me to some material that explains submit processing, redirects, etc. I would be happy to do some more digging.
Thanks greatly in advance.

Thanks Scott,
I may be being dumb here but I don't see how that would help...
P250_PROJECT and G_PROJECT are currently kept in sync by app logic. Whichever is used to drive, if the page is rendered from cache then the app logic is not executed, so the rendered page contents are not those keyed by P250_PROJECT, as illustrated in steps 1-4 of the problem above.
The user sees X, the session items say Y. The engine doesn't know what the user is seeing.
when page P is POSTed, its hidden item P250_PROJECT should always be used to derive the application item G_PROJECT. Then whether the page was pulled from cache or rendered anew via a click from the sidebar link, the project ID is determined by the contents of that page.
As I said above I tried this, with the 'Edit' branch set to:
Set these items: G_PROJECT
With these values: &P250_PROJECT.
but it makes no difference. The project ID is not determined by the rendered page contents - the engine gets the value of P250_PROJECT from session state.
I can code the 'Edit' pages such that they check permissions and if necessary redirect back to p250 (conditional before-header branch), but that's a clunky cure rather than the prevention I was hoping for.
Please tell me if my understanding is incorrect.
jd

Mapviewer javabean, sessions, state and related

Hi,
i am a scratching my head a little bit regarding the way I use MapViewer.
We are building a web-app using a mapviewer for rendering the map, doing queries etc. We are using AJAX for user-interaction, including the map window, identifying objects, searches etc.
I have currently implemented a singleton class which holds one instance of MapViewer. This instance recieves parameters from the session, and then executes run();
However, since the MapViewer bean is stateful, and I want to make use of identify() etc, I realize that this is probably a wrong way of doing it ?
Would it me more appropriate to alloc a new MapViewer instance for each session ? And would this be a very memory-consuming way ?
kind regards,
Oyvind

Thanks, Andrejus.
I tried to do some memorytesting, with results:
100 instances:
com.powel.mapviewer.MapViewerFactory produced [Ljava.lang.Object; which took 1403184 bytes
1000 instances:
com.powel.mapviewer.MapViewerFactory produced [Ljava.lang.Object; which took 14086016 bytes
10000 instances:
Exception in thread "main" java.lang.OutOfMemoryError: Java heap space
Seems like this is a memory-eating way of doing it, so I think I will try to create a pool of mapviewers instead.
Oracle Maps is an interesting feature, but in our case I believe it will complicate things (we dont want tiles, and we use WMS background etc)
Regards,
Oyvind

Session state and LOVs

Hello,
I'm pulling my hair out attempting to have an LOV act correctly (like an unruly kid). The problem is where you are selecting an existing record and then this is taking you to an update page. On the update page, there is LOV1 where the value should already be set from the page query (which pulls-in all the other data for the page, not just the LOV value). This value is then used to determine the contents of LOV2. However, LOV2 is always null (empty) until I actually choose something different from the existing value in LOV1, forcing a redirection and then choosing the original value again (to set everything right).
Why does my session values not seem to be set? This is preventing me from letting the testers test this app before release. Any insights? Any workarounds? All suggestions welcome.
thanks

Jim,
Have you tired using the Page Processing in which you can assign your values to be fired throughout the application?
You assign the value based on your original LoV and then you can pass these values (copy the computations to other pages).
Note: In doing so you will need to ensure you are addressing the original item and page. With this also ensure that you maintain continuity of item names across the pages and that on each new page which uses the original LoV (item) that you call this item out in the source value or expression.
Randy

Row fetching, session state, multi-page processing of 1 row - problems

I'm building an application that's updating a couple of tables via an updateable view and an instead-of trigger. That portion of it works OK, EXCEPT for the problems that I'm having with the APEX end.
I've got 3 pages; the first page reads the row in via the automated row fetch and has, say columns 1-20.
Page 2 has, say, columns 20-40 and page 3 has the remainder.
I have the row fetch set up conditionally so that it only executes the first time that page 1 is displayed. All the columns are set as Database Columns.
It seems that if I fill in fields on page 1, move to page 2, and back to page 1, the contents of the fields on page 1 are destroyed unless I set their 'Source used' to 'Only when the current value in session state is null'.
Furthermore, upon update (which is on page 3, via an automated row update), the only columns that get updated are those that exist on page 3. The column values from the other pages no longer exist.
Does navigation to another page not retain the session-state of database-column items?
Am I better off just writing my own processes to perform the updates?
Any other suggestions?
Frank

Hello Frank,
>>I have the row fetch set up conditionally so
that it only executes the first time that page 1 is
displayed …
It seems that if I fill in fields on page 1, moveto page 2, and back to page 1, the contents of the
fields on page 1 are destroyed unless I set their
'Source used' to 'Only when the current value in
session state is null'.
It seems that the first sentence is not consistent
with the second one. The only reason the values are
destroyed, as you call it, is that the ARF is running
again. What condition are you using on your ARF?
In any case, the right way to deal with it is to
change the source used to "Only …", as I understand
you already did.
Hi Arie,
I set an item after the first ARF. I then have the ARF set to execute conditionally based on the presence of that item. (I also display a success message on the ARF, so I know that the ARF is only occurring once). I move to subsequent pages by submitting the current pages.
>>Does navigation to another page not retain the
session-state of database-column items?
Navigating to another page does retain the session state values.
It's depends on what you mean by navigating.
Submitting the page set/update session state.
Redirect without submitting the page will not
set/update the session state.
>>Furthermore, upon update (which is on page 3,
via an automated row update), the only columns that
get updated are those that exist on page 3. The
column values from the other pages no longer
exist.
Automatic DML is working only with the current page
items. You can't use it to insert/update other page
items, even if they are in session state. However, in
your case, it's not the Automatic DML performing the
actual updates, but your triggers. Are your triggers
relying on session state values?
I think this statement 'Automatic DML is working only with the current page
items.' answers my question. My trigger is working on session state values. It's an 'instead-of' trigger, so I was letting the automatic DML perform the 'update' on the view, while allowing the trigger to do the actual update of several tables that the view joins by referring to their :NEW. values in the trigger. Since the DML occurs on my last page, those were the only session state values that my trigger ultimately saw.
>>Am I better off just writing my own processes to
perform the updates?
I believe the correct way is to submit the page and
then branch to the next one. Submitting the page will
save all your items in session state, and allow these
items to be available for page 3 DML.
Regards,
Arie.I'm doing that already - all the values do appear to be in session sate, except only the items that are on the last page actually make it to the automatic DML update, which seems to agree with your previous statement.
I suppose I could write my own process to perform the update, as the values are available in session state.
Thanks,
Frank

Session State Values Only Used After a Page Refresh

We recently moved our application to use Oracle Application Server. I don't know if this is an issue but I thought it might be relevant to my question.
In this application there is report that has a link to another page. This link sets the ID of the company that the subsequent pages use to provide several summarized reports with drill downs. I'll call this the main page.
Users recently started to report that when they drill into the details and use the breadcrumbs to get back to the main page, that page will display the wrong company, one they were viewing recently.
In testing I find that when I click a link to drill down into the details the details page is not for the right company. I check session state and the value of the ID is correct. If I refresh the page I then get the correct drill down page. Some of the drill down reports will pass a page specific ID into each page and others will just reference the ID from the main page. Not sure if this is relevant since the breadcrumbs to go back to the main page would not be able to use every possible page specific ID, my understanding was that once I set the ID in session state on the main page I could rely on it.
Also if I get the wrong details page and click on the breadcrumb to the main page I get the wrong main page (it displays a previously viewed company). Again the session value is correct and a refresh of the page will bring back the right company.
Our systems admin was asked to check if we had WebCache on and she said it was not. Is there something else that could be caching the results of these pages?
Thanks,
Greg

Scott,
This is in version 2.2.1 and there are no caching features available.
The application does require login.
I'm playing around with the APP_UNIQUE_PAGE_ID right now but I am finding that even in the builder if I edit the attributes of a report column and then try to edit another column it will bring up the last column I edited. Even if I use the record navigation buttons next to the Apply Changes button it will keep bringing me the same page over and over unless I constantly refresh the pages.
Greg

Element session state Question

Hi
I have built a parametrized report using an Element P4_JOB and by clicking on a Button, Some Employees Data should be displayed in a Region.
my Problem is i want to reset the Value for the Element P4_JOB when the Buton is clicked and the result is diplayed.
how to achieve that?
br.
MDK.

I think you'll need to clear the item when the page is being rendered to show the results. If you try to clear it earlier it won't get set into session state and be picked up by the reporting query.
You could set the source for the item to be static text and leave the source empty, and also it so that it resets the item's contents on each page load rather than when it's value is null (sorry I can't remember the exact wording on the apex builder)

Custom Tabular form - Session state values disappear

Hi,
I have a custom tabular form page using apex_item and apex_collections. I am using the apex_collection to store the values in memory in order to avoid the data loss in session state. The page is working fine after a validation failure and it shows the values on the second report (as advised for custom tabular forms).
In this, some of the rows have clob value which needs to be edit on a rich text editor. Since apex_item does not provide a solution to have rich text editor, I have to branch to another page on an Edit link for the CLOB data using a modal window. However When I return back from the modal page, the values disappear from session state, and the validation (using htmldb_application.g_fxx arrays) does not take place.
Is there any workaround to overcome this issue?
Apex version: 4.1
Thanks in advance.
Natarajan

Thanks Daniel for your reply.
Actually its an interface with a custom tabular form on page 1 with apex collections and rich text editor on page 2. Since it is not possible to have the rich text editor, I need to have it in another page (I could have it the same page, that is the way I am trying now) however I open the next page as a modal window.
I have a link on the tabular form, so that the clob content is opened in the rich text editor on the next page. After the edit, I update it to the same collection and go back to the tabular form page and refreshes the report. It works fine, but causes trouble to the data entered on the other columns in the custom tabular form.
I have now deviated to another work and put this on a hold. If my explanation is still not clear, I will create the same page in apex.com to simulate the error.
Thanks again.
Natarajan
Edited by: Nattu on Mar 14, 2012 11:35 PM

Htmldb_Get() not updating session state?

Thankfully, I solved this problem myself before posting this, but I still have questions about why I ran into the problem in the first place.
Alright, I have a javascript controlled interface on an ApEx page that uses htmldb_Get() for retrieving the current state of a row in the database and for submitting the data back to a page process to update the database row. I've checked my values from the javascript function that saves the data and the strings do contain the updated text from the interface, but when my page process fires off the strings have reverted to what was in the session state from the earlier AJAX request when the interface was loaded.
* I've checked the session state and the variables in question are reverted, but I noticed that the item statuses are listed as "Updated".
* I've also tested this process by passing the request string and page items in the URL and the process functions correctly.
* I've double checked the item definitions to see if the items were being overwritten (they weren't).
* I've checked the item types to make sure that none were "protected" (they weren't).
* I haven't found any error messages in logs within ApEx.
At this point I noticed that the AJAX requests to load the interface were showing up in the "Recent Page Views" report, but the save requests were not. For no particular reason I decided to switch from using the request string in the conditional for the page process and use a page item to control the mode as so...
WAS:
var get = new htmldb_Get(null,$x('pFlowId').value,'save_asset_tracking_info',11);
NOW:
var get = new htmldb_Get(null,$x('pFlowId').value,'',11);
get.add('P11_FORM_MODE','save_asset_tracking_info');
...and magically the page items in the session state are being updated and all is well. There are few things more frustrating than spending hours debugging a process, then fixing it with an action that you are completely convinced won't change a thing.
This make no sense to me and I'm hoping someone can explain it or at least point me in the right direction.
Thanks in advance,
Jason

Apologies if I skipped a few details in my first post...
I haven't included much of my source code because I don't believe it's the key to this issue. If you disagree, I will post the source code for your reference.
On page 10 of my application, I have a floating div interface that appears in front of the page content. This div is initially empty, but I use JavaScript to make an AJAX call to page 11, strip out a specific region using the get.get() function. On the floating interface, there is a "save" button that fires off another JavaScript function that uses htmldb_Get() and get.add() to post all 21 of the items from the floating interface back to a page process on page 11. The "save" page process is at process point "on load - before header" and is extremely simple -- "if ID found then update record, else insert record".
The problem I had was that after the "save" button was pressed, all of the items sent via AJAX to the page process were always identical to the initial values from when page 11 was rendered and the floating div interface was populated. This seemed like some sort of quirk or issue with the session state cache overwriting the values that were being sent via AJAX. As I said before, I checked the values in JavaScript immediately before the AJAX "save" submission and everything was kosher, but the page process never got those updated values.
I managed to eliminate this "symptom" by simply...
1. Setting the htmldb_Get() request string to null ("")
2. Adding one more get.add() call for P11_FORM_MODE
3. Changing the condition on which the page process was fired from the request string to the page item
All of the other get.add() calls remained the same as did the PL/SQL in the page process.
Thanks,
Jason

Problem with application item and session state

Okay, let's see if I can explain this problem coherently.
I have a small app (one page), with an application item, F_WHERE_CLAUSE.
This page has three regions in which there are items that the users can populate for search conditions. A couple of these items are "select list with submit" (I still need to upgrade to the AJAX method, I know). There is another region which has one hidden field, called P1_WHERE_CLAUSE. This field is defined to "Always, replacing any value in session state..." with source type of "Item (application or page.....", and a source value of F_WHERE_CLAUSE with no default value.
I have a button called "Search" which submits the page and fires a PL/SQL process which builds a where condition based upon the other page items and stores the value to the application item F_WHERE_CLAUSE (correctly).
For testing, I've made the P1_WHERE_CLAUSE field visible so that I can see what's going on. I've also clicked the debug and session buttons to help trace this. After I click the "Search" button and the page submits, debug shows:
0.02: ...Session State: Save "P1_WHERE_CLAUSE" - saving same value: "1=1"
followed later by:
0.05: ...Session State: Saved Item "F_WHERE_CLAUSE" New Value="lower(primary_class) = 'rock' and country = 'Spain'"
The field P1_WHERE_CLAUSE displays with the correct search criteria as signified by F_WHERE_CLAUSE above. However, If I click the "session" button to view the session state values, P1_WHERE_CLAUSE shows up as:
P1_WHERE_CLAUSE Textarea 1=1 U while F_WHERE_CLAUSE displays the correct value still.
The reason this "problem" came up, is that this page also has three SQL report regions which use &P1_WHERE_CLAUSE. for the where condition. While they display the correct results on-screen, each report region also has the "Export to csv" enabled, and the export seems to be using the "1=1" condition (from the "session" window) instead of the search criteria that the on-screen region is using (F_WHERE_CLAUSE and the displayed P1_WHERE_CLAUSE), resulting in a retreival of all records.
Anybody have any idea what's going on and why, and how to get the csv export to use the correct value for the where condition?
Thanks,
Bill Ferguson

It appears the "Export to CSV" functionality requires the item value to be set in session state. The P1_WHERE_CLAUSE item value never gets saved to session state. The page is rendered and the value is put in the item on the page but until you submit the page session state doesn't know what P1_WHERE_CLAUSE is.
Create a before header computation or process to set the value of P1_WHERE_CLAUSE (which will save it to session state). It is interesting that the report regions didn't need to look at the value in session state but the "export to csv" does.
--Jeff

Report Print Attributes and Session State Protection

Hi all,
I have a report (old style Apex report, not an IR) that I would like to have a Print option, producing a PDF output.
The report is on a page that takes a number of arguments and uses the page security option "Arguments require checksum".
When I turn on the Enable Report Printing option in the report, the Print link appears at the bottom of the report. However, clicking Print leads to the following message:
"Error No checksum was provided to show processing for a page that requires a checksum when one or more request, clear cache, or argument values are passed as parameters."
Disabling Session State Protection for that page makes the report link work, but I would prefer not to do that.
How do I get the built-in report print option to generate the required checksum?
I'm using Apex 4.0.2.00.07, by the way.
Alex

WORKAROUND:
1.) Create hidden ITEM on page (I named it P23_PREPARED_CSV_DOWNLOAD_URL).
Enter the following for the ITEM
as the SOURCE_TYPE : PL/SQL Function Body
as the SOURCE: return apex_util.prepare_url('f?p=&APP_ID.:&APP_PAGE_ID.:&APP_SESSION.:CSV:')
2.) Create BUTTON that executes javascript to open POPUP window with this url.
a.) Create Button and enter
<a href="javascript:popupURL('&P23_PREPARED_CSV_DOWNLOAD_URL.')">Download and Save to CSV file</a>as the "Text Label/Alt"
Originally had custom code for javascript POPUP and this is not needed .... just use the APEX javascript function.
Edited by: Bryce Tuohy on Mar 5, 2009 10:47 AM

Interactive Report Download and Session State Protection

I have created an Interactive Report in an APEX application that I have enabled
session state protection for. The issue I am having is with the "Download"
functionality of the interactive report to a .csv file.
The URL created by selecting Download from the drop down (javascript:gReport.controls.download();)
is built or constructed without a checksum thus causing the error below.
Error
No checksum was provided to show processing for a page that
requires a checksum when one or more request, clear cache, or argument
values are passed as parameters.
The anchor tag containing the URL (/f?p=app_id:page:session_id:CSV:) is contained within
<div id="apexir_CONTROL_PANEL_DROP" class="drop_panel
clearfix" style="">
Running the following: Application Express 3.1.0.00.32 on Oracle Database 10g Enterprise Edition Release 10.2.0.3.0
Is there a way to add a checksum to this? OR does anyone have any ideas on how to work around this?
Edited by: Bryce Tuohy on Feb 26, 2009 10:08 AM

WORKAROUND:
1.) Create hidden ITEM on page (I named it P23_PREPARED_CSV_DOWNLOAD_URL).
Enter the following for the ITEM
as the SOURCE_TYPE : PL/SQL Function Body
as the SOURCE: return apex_util.prepare_url('f?p=&APP_ID.:&APP_PAGE_ID.:&APP_SESSION.:CSV:')
2.) Create BUTTON that executes javascript to open POPUP window with this url.
a.) Create Button and enter
<a href="javascript:popupURL('&P23_PREPARED_CSV_DOWNLOAD_URL.')">Download and Save to CSV file</a>as the "Text Label/Alt"
Originally had custom code for javascript POPUP and this is not needed .... just use the APEX javascript function.
Edited by: Bryce Tuohy on Mar 5, 2009 10:47 AM

Session pooling and statement handles

Hi there,
I have a large multi-threaded application (perhaps >100 threads). Each thread is continuously processing events (very high volumes) which involves some manipulation and some database operations (from a fixed set of possible operations).
I am using session pooling but what I want to know is, Should I:
(a) Prepare my fixed set of statement handles up front at program start-up when I'm creating the session pool and then reuse the statement handles in each event processing thread (also, is this thread safe ? even if it is, all threads would be contending on the same statement handles)
or
(b) Prepare the statement handle for each event which presumably will exploit the statement cache on the session pool. This would also mean not having statement handles shared between threads thus removing any thread contention issues.
I think (b) is the option for me, but does anyone have any thoughts ?

With a), one would think it's OK, but I would hate to find out that it's not thread safe by accident.
But anyway, with b) the cost of allocating private statement handles in each thread seems very low. The memory required for the statement handle plus its bind and define handles could very well be below 8k per statement. If you've got say, 5 statements * 100 threads, you're only looking at around 4MB overall.
Finally, you might want to make sure that the session pool statement cache is working by checking the values for 'executions' and 'parse_calls' in V$SQL for your statements.

Session state and PDF building

Similar Messages

Maybe you are looking for