Some months ago I published an Oracle Apex plugin to restrict the left side shuttle choices. You can read the original post here.
On the demo page the On Change event was used to refresh the left side list. My friend Alex Nuijten made the remark this was not very intuitive. The user needs to navigate out of the selection item to see the change.
So I changed the demo page to provide instant feedback. The selection is now refreshed after each keystroke. As this process is completely executed on the client the response is very fast. Only for very long list the user might notice a slight delay.
No changes to the plugin were needed. It can still be downloaded at

Here is how is was done:

Selection item
Type: Text Field

Dynamic Action
Name: Select on keystroke
Event: Key Release
Selection Type : Item
Item(s): P400_SELECTION

True Action
Name: Perform Selection
Action: Restrict Shuttle Values [plug-in]
Search Item: P400_SELECTION
Selection Type: Item(s)
Item(s): P400_SHUTTLE
Fire on initialization: No

Happy Apexing

Source Article from

A few weeks ago I wrote a post how to avoid multiple logins when using a link in an email. This solution however does not work in a batch process for two reasons:

  • the URL needs to be fully specified, so protocol, server, port and directory are also needed. These data are not available outside the Apex context
  • the function apex_util.prepare_url can only be called from within an Apex session

My first idea was to prepare the URL for every possible mail while being in an Apex session and store it in the database. The stored URL can be used when generating the email.
This would be technically  complex, because a new column would have to be created and the content of this column would have to be kept up to date all the time.

Another possibility would be to create a fake Apex session within the batch process. Then the rest of the code need not be changed. Martin Giffy d’Souza provides a way to create an Apex session in this blogpost. Although this is an 5 year old post I was able to create an Apex session and generate a valid link.

The last solution was chosen because of the simplicity and the absence of need for code change.

Happy Apexing!

Source Article from

Some Apex applications sends notifications emails to users when tasks are due. These emails contain a link to an Apex page. When the user clicks on the link he is taken to the page. When this page is not public – which usually is the case – he is directed to the login page of the application. Even when he is logged in for the same browser on another tab.
When users get a lot of these emails it is annoying they have to log in each time. Apex provides a way to avoid the login when a session exists in the same browser. This post describes how you can set it up using the Rejoin Sessions functionality.

Source Article from

For the Apex Dashboard Competition of the DOAG in 2016 I created a Apex dashboard. You can read about it in this post
One of the elements of this dashboard was a group of four infocards.

These cards display the key values for a country on a certain subject. Apart from the text the cards differ in color and the icon used.
These cards are created using a named report template. This is a special kind of report template. Using a normal report template the query columns and rows map to the columns and rows in a HTML table. A named report template consists of a HTML template in which the query columns are referenced by name. For the infocards the query and template below are combined:

select 'Population'  title
     , population    data
     , 'Number of inhabitants'  text
     , 'fa-users'   as  icon_class
     , 'db2_red'     as  container_class
from   cnt

Row Template 1 within the Report Template:

The substitution strings in the template are replaced with the corresponding values from the query.
As you can see the column values can also be used to define CSS classes as in the enclosing DIV element. The value of the column container_class defines the name of the class to applied. This class defines the color of the card. The class icon_class defines the Font Awesome icon to be used.

The cards are styled with the following CSS:

The color of the card can be defined by selecting the db2_color class as container_class in the query. 

The icon is positioned absolute relative to the card with the class db2_icon_container. Space is created for the icon container by defining  left:80px; for the main DIV containing the text.

The rest is quite straight forward CSS styling.

You can create your own Named Report Template:
– navigate to Shared Components > Templates
– press Create
– chose Report
– create the template From Scratch
– enter the name for your template and check Named Column (row template)
– then enter the HTML for the template

Named Report Templates provides us developers with more freedom to style the output. The downside is that the templates are more specific and less widely applicable. 

Happy Apexing

Source Article from

The Oracle Apex shuttle item is a neat way to select a number of values. It is however not so user friendly when the list of choices is very long. In this case it is useful to be able to limit the list of choices.
When you implement the limitation in the LOV-query of the item, the limitation will also be applied to the selected values. Usually this is not the desired behavior.
This behavior is avoided when the restricting of the left shuttle pane using JavaScript. I have created a Dynamic Action plug-in to do this. The choices in the left shuttle pane are evaluated case insensitive against the content of the search item. If the search item contains more than one string ( separated by spaces) all the strings should occur in the shuttle value. The search string ‘INVOICE 2013’ returns only values that contain ‘INVOICE’ (independent of case) and ‘2013’.

The plugin attributes are the shuttle item and the select item. The plugin needs to know the select item to disable submit when enter is pressed.

You can see the plug-in in action on:

As usually you can find the plugin on

Happy apexing

Source Article from

Working with Apex I regularly write Javascript. Most of the time these client-side code snippet are blazing fast, but in some cases they take a few seconds to execute.

In these cases I want to see which part of the code is slow. In this process I use two very simple JS functions which make life a lot easier for me:

var timing_start;

function start_timing( text) {
    timing_start =;
    if ( text) { console.log('Start timing:'+text); } 

function show_timing(text)
{ console.log('Timing:'+text+', milliseconds:'+ ( );  }

You can put this code on the page or include it in a general JS file.

 The use of these functions is really simple:

code to be timed 
show_timing('Data retrieved');

This will result in the following output in the browser’s console:

Timing:Data retrieved, milliseconds:784.2350000000001

The show timing calls display the elapsed time since the last call to start_timing.

Happy JavaScripting

Source Article from

Recently I’ve been installing the OATOA (Oracle APEX on Tomcat with ORDS behind Apache) stack a number of times on both Oracle Linux 7 and CentOS7.

In the past this was no problem, Tomcat came up and a request to http://myserver:8080/ords/ gave me the expected APEX login screen.

However since Oracle Linux 7 and CentOS7, it seemed that Tomcat started before it could find the Oracle database. A simple restart of the tomcat service would do the trick. On my virtual machine (recently switched from VirtualBox to parallels btw) meant for doing presentations no problem at all.

systemctl restart tomcat

Now I’m playing around with “droplets” at DigitalOcean, just another name for virtual servers in my opinion, but they have great features at a reasonable price. Take a look:

But my droplets showed the same behaviour as my parallels virtual machine. I first had to restart tomcat to get a working system.

You all should know that I’m not a linux guru (really I’m not), so broke my brains about this for a looooong time.

Yesterday I finally managed to get this working. The idea is to make the tomcat start-script wait until it can see “something” on port 1521 (the db listener) and only then continue to start.

Open the file /usr/libexec/tomcat/server

nano /usr/libexec/tomcat/server

now add some lines of code _before_ the line that starts with MAIN_CLASS like this:

. /usr/libexec/tomcat/preamble
# .=.=.=.=.= START Make tomcat wait on oracle .=.=.=.=.=
while netstat -lnt | awk ‘$4 ~ /:1521$/ {exit 1}’0
  sleep 10
  let i+=1
  if [ “$i” -gt “5” ]
    break    #Abandon the loop.
# give Oracle some slack to also start the database
sleep 10
# .=.=.=.=.= END Make tomcat wait on oracle .=.=.=.=.=


add the obvious lines to your script.

What is happening?

The line “netstat -lnt | awk …” tries to find the string “:1521” in the result of a “netstat -lnt” command. This would indicate that something is listening on that port. In our case it will be the Oracle Listener.

If it does not find an open port 1521, it will sleep for 10 seconds, increment a counter (“i”) by 1 and do the loop again. The loop will loop a maximum of five times to prevent an endless loop.

After the loop completes we give the Oracle database another 10 seconds to get started.

That should do the trick. It does for me (I use the same code at digitalocean as well as my parallels virtual linux server.

I’m sure linux must have some system in place that could do the trick as well (make one service dependable on the other) but I haven’t found an easy one to use. If someone could help me out here, feel free to leave a comment.



One of the things that cost me a lot of time is the writing of log statements.
I do it a lot when I am working on complex PL/SQL or JavaScript routines. Just to understand what’s happening. Mostly these log statements have the form:

logger.log(‘varname1=’||varname1||’, varname2=’||varname2, etc… );

and most of the time the variable names are even longer :-(. 
So it is a lot of typing, especially for something that is deleted after a few hours. As I am into automating my job I decided to automate it. Here you can see it in action: 

The selection and generation is all written in JavaScript. You can see how fast it works ( it is recorded in real time ). The code is pasted in the textarea, variables are selected by clicking and then the log statement(s) can be generated. 

You can find the page here:

The page supports PL/SQL and JavaScript. For PL/SQL the identifier naming rules are quite strict, so these are implemented. At the moment the same rules are also applied to JavaScript. 
NB As JavaScript rules are very free, some variables might not be recognized, for example when they contain diacritics.

For PL/SQL you can select logger or apex_debug, for JavaScript console.log statements are generated. 
There is an option to generate mulitple calls or combine all variables in one call. 

I have tested this with PL/SQL and JavaScript and I could run the generated code as is, no changes were needed. 

Happy Apexing and JavasScripting

Source Article from