Thursday, 3 July 2014

Oracle vim config

This page gives a pretty good overview of tweaks you can do, so I just want to re-gurgitate what I have done (from that site).

Syntax Highlighting

Grab the file:, and save this in: ~/.vim/syntax/

Then edit your .vimrc file and add the following line (to auto detect the filetype based on the extension):

au BufNewFile,BufRead *.fun,*.pks,*.pkb,*.sql,*.pls,*.plsql    set filetype=plsql

Auto Indentation

Grab the file:, and save this in ~/.vim/indent/

Then edit your .vimrc file and add the following line

runtime! indent.vim


Though, I prefer just to use auto indent - there are some special situations where it wont calculate the indentation correctly using the indentation file above.

So this gives me the following in my .vimrc file:
au BufNewFile,BufRead *.fun,*.pks,*.pkb,*.sql,*.pls,*.plsql    set filetype=plsql
au BufNewFile,BufRead *.fun,*.pks,*.pkb,*.sql,*.pls,*.plsql    set nosmartindent

Wednesday, 2 July 2014

Accessing Google data, part 2


Back in Novemeber 2012, I blogged about accessing Google data from APEX - and provided some sample code to getting started. You can find this post here:

Over the past little while, I have been (slowly) working on a more modularised API, that can be easily built upon to add support for additional Google services. I have called it PL/GAPI (Google API). It can be found on github:

I hope the wiki on github is straightforward enough to follow in setting this up in your local environment:


I've started with the following services:
  • Drive
  • Calendar

Without going into too much detail (since it's all on the wiki linked earlier), you link to an authorization URL with: 


This tells the system where to return to and P0_SCOPE defines what scope you are requesting authorization for. Current scopes defined:

Once the user accepts, they are returned to the application in the parameters specified in the above redirect (p_return_app, p_return_page) along with the item to store the refresh token in.

Refresh token is a means to be able to get the access token without continually prompting the user for access. All requests require a valid access token to complete successfully. Because the returned refresh token contains the '/' character, it is escaped by APEX. Before storing it, you need to un-escape the string. To do this, I used the function: UTL_I18N.UNESCAPE_REFERENCE. 

Once we have the refresh token, we can easily get the access token with:


Read more about the implementation here:

Then for example, we can add a new folder to our drive:

            p_folder_name       => :P2_NEW_FOLDER_NAME
          , p_access_token      => :GOOGLE_ACCESS_TOKEN);

I've set up a few examples on the wiki, but really, at this stage, the sample application in the samples folder is the best tool to refer to for sample usage


Any feedback, criticisms, suggestions about the implementation are more than welcomed.

note: It's a long way from being complete, but just thought it was time to share.

Saturday, 3 May 2014

Triggering dynamic actions from a dialog button

A typical usage scenario is you create a region, give it a static ID, and set the default display to be hidden - the latter is handled if you select your region template as modal region.

A typical JavaScript dynamic action to display the dialog, would be:

    width: 500,
    title: 'My Dialog',
    buttons: [
            text: 'Submit',
            click: function() { 
} }, { text: 'Cancel', click: function() { $(this).dialog("close"); } } ] })

So the question is, how do we get some other dynamic action to fire when the button is clicked?

Option 1

On the button, give it an id property, such as:

    text: 'Submit',
    id: 'btn-submitDialog'
    click: function() { 


And then, on your dynamic action definition, you can specify a click event and on the selection type specify jQuery Selector, passing in #<id>. So given the above button, you would pass in #btn-submitDialog. Also, specify the event scope as dynamic.

Option 2

Trigger a custom event. This option has the benefit that if you want the same dynamic action to fire from multiple sources, just trigger the same event.

On the button click handler, fire apex.event.trigger(document, 'btn-submitDialogClicked'), such as:

    text: 'Submit',
    click: function() { 
        apex.event.trigger(document, 'btn-submitDialogClicked');


Then, create a dynamic action specifying the When as Custom (under the Custom Event heading). As from the triggered event above, specify the string: btn-submitDialogClicked. Specify Selection Type as DOM Object, and the DOM Object as document.

Option 3

Craft the display of the dialog based on a before page submit dynamic action adding a when condition such that it only appears when the field is empty.

E.g. You may want to display a dialog on certain conditions when the user submits the page.

In this case, the click handler would just submit the page

    text: 'Submit',
    click: function() { 


Then add a dynamic action with the event being Before Page Submit (under Framework Events heading). In the when condition, specify on what situations the dialog should appear. So in my sample I have a field P21_SAMPLE_FIELD1 that I want a value in before actually submitting the page, so I specify the when condition as JavaScript expression, and the value: $v('P21_SAMPLE_FIELD1') == '' && == 'SUBMIT'

Then you want to add two true actions:

1. An Execute JavaScript Code dynamic action that will render the dynamic action (as described at the start of the article)
2. A Cancel Event action (under the Miscellaneous heading) so the page doesn't get submitted. Note: you will want to have this sequenced after displaying the dialog, otherwise the dialog won't appear.

nb: Option 3 isn't exactly firing another dynamic action when clicking the button, but continues submitting the page when the action's condition is not met - I hope it gives you an idea of some options in regards to page processing and dialogs.

I have set up a basic demo with the 3 options in force on the following page:

Thursday, 17 April 2014

Times at specific time zones

These past couple of weeks, I have been doing some work with some external APIs, so have had to do some time stamp manipulation. Here are some tips I've learnt along the way.

A quick way to get UTC time, is with the function: sys_extract_utc. With that, we can quickly get the UTC timestamp. Here is an example to return the UTC time in RFC3399/ISO8601 format:

  , 'yyyy-mm-dd"T"hh24:mi:ss.ff3"Z"'

To return that back into local time, you want to declare a variable with time zone support. So just say the string you have is: 2014-04-17T02:46:16.607Z, you'll want to declare it with a time zone attribute.

l_time timestamp with time zone;

Since Z refers +00:00, you need to set the time zone. If you call to_timezone or to_timezone_tz, without specifying the timezone, it will create the time in the local time zone, so we need to specifically set it with from_tz.


l_time := to_timestamp_tz('2014-04-17T02:46:16.607Z' || '+00:00', 'yyyy-mm-dd"T""Z"tzh:tzm');

(Appending the date time string with the time zone info)


l_time := to_timestamp('2014-04-17T02:46:16.607Z', 'yyyy-mm-dd"T""Z"');
l_time := from_tz(l_time, '+00:00');

(Setting the time zone after the fact)

Finally, to get the time stamp outputting in your local time zone, you can do:

l_time := l_time at local;

(Converting it to the time zone of the database)

Or, specifically set what time zone to convert it to:

l_time := l_time at time zone '+10:00';
l_time := l_time at time zone sessiontimezone;
l_time := l_time at time zone 'Australia/Sydney';

Where the time zone string can be a valid time zone name as from the v$timezone_names view; or the GMT offset, as per the example above.

If all you want to do is get it in the local time zone of the database, you can declare the time stamp with local time zone, and this will automatically output the time in the databases local time zone.

  l_timestamp timestamp with local time zone;

  l_timestamp := to_timestamp('2014-04-17T02:46:16.607Z', 'yyyy-mm-dd"T"hh24:mi:ss.ff3"Z"');
  l_timestamp := from_tz(l_timestamp, '+00:00');

  dbms_output.put_line(l_timestamp );


This seems like as good a time as any to spruik the fact that Application Express comes with support for local time stamps since version 4. You can read more about that here:


Thursday, 3 April 2014

Getting around large exports with RTF report templates

If you are on a slow connection like me, you may find how much of a pain it is deploying apps that house some BI publisher report templates. It only takes a few templates before your app export becomes 10MB+ in size.

Perhaps in a future release, report templates will be treated in the same way as application images, in that you can add them as a script to supporting objects so that you don't need to export the templates each and every deployment.

If you haven't seen, apex_util has a procedure to download report queries, so by using that procedure coupled with storing your files in your own maintained table, you can avoid the issue of massive application exports.

First create a table to store the templates:

create table rtf_template(
    code varchar2(25) PRIMARY KEY
  , layout BLOB);

Then, add your template file to the table.

We will need two application items (or 1 will do if you identify the layout with exactly the same name as the query). So I've created: REPORT_NAME and REPORT_LAYOUT

Then create an on demand process, which I've named mine as: PRINT_DOC. The code looks like this:


    l_file_as_clob CLOB ;
    l_layout BLOB;
    l_asdsad CLOB;

    l_dest_offset integer := 1;
    l_src_offset integer := 1;
    l_lang_context integer := dbms_lob.default_lang_ctx;
    l_warning integer;

    select layout into l_layout
    from rtf_template
    where code = :REPORT_LAYOUT;
    dbms_lob.createtemporary(lob_loc => l_file_as_clob, cache => false);
        dest_lob => l_file_as_clob
      , src_blob => l_layout
      , amount => dbms_lob.lobmaxsize
      , dest_offset => l_dest_offset
      , src_offset => l_src_offset
      , blob_csid => dbms_lob.default_csid
      , lang_context => l_lang_context
      , warning => l_warning);

    l_file_as_clob:= replace(l_file_as_clob,'+','%2B');
    l_file_as_clob:= replace(l_file_as_clob,'/','%2F');
    l_file_as_clob:= replace(l_file_as_clob,'=','%3D');

    --signature 2
    apex_util.download_print_document (
        p_file_name => :REPORT_NAME
      , p_content_disposition => 'ATTACHMENT'
      , p_application_id => :APP_ID
      , p_report_query_name => :REPORT_NAME
      , p_report_layout => l_file_as_clob
      , p_report_layout_type => 'rtf'
      , p_document_format => 'pdf'

Now, when ever you want a button to download a layout, you can just past the request parameter of: APPLICATION_PROCESS=PRINT_DOC, then set the two (or one) page items accordingly. A little like this:

It is a little extra work, but if it saves some headaches with deployments, why not ;-)

Monday, 31 March 2014

Database connectivity with Google Sheets


I happened to be watching some Google I/O videos from 2013 last week and was watching some on apps-script, which I had never really used up until now. They mentioned about having JDBC connector so that you could connect to external data sources (Oracle, MySQL, SQL Server, etc).

It's worth noting, there are 2 types scripts that can be container bound to a spreadsheet:

1. Add ons
2. Script gallery

Script gallery is not supported in the new version of sheets, in favour of add-ons. There is a market place where you can search for and install add-ons, and if your organisation is using google apps, you can even restrict your add-on to your organisation.

Unlike the script gallery where you could view the source, that is not the case for Add-ons (the code is not visible to end users). Excluding if you share the document with the development version of the code attached to the document.

Add-ons are currently in developer preview, so whilst you can develop add-ons, only approved developers can publish to the store.

For more information on apps-script in general, head over to

Database Connectivity

All the information is on their JDBC guide:, all the same, here's a little demo of connecting to an Oracle database and populating cells. I will be fetching some data from my app_users table.

Open the script editor by opening the sheet and navigating to tools --> script editor. The name you give the project is what will appear in the Add-ons menu, and then any menu entries you create will appear as a sub menu-item.

The first step is to add some menu entries:

function onOpen(e){

  .addItem("Fetch app_users", "loadAppUsers")
  .addItem("Clear sheet", "clearSheet")

function onInstall(e){

The onInstall function is there so the menu is added when the add-on is first installed, with onOpen being called whenever the sheet is opened.

Then a function to connect to the database and populate the spreadsheet:

function getConnection_(server, sid, port, username, password) {
  return Jdbc.getConnection('jdbc:oracle:thin:@//' + server + ':' + port + '/' + sid, username, password);

function loadAppUsers(){
  var conn = getConnection_(
  , '[sid]'
  , '[port]'
  , '[username]'
  , '[password]'
  var stmt = conn.createStatement();
  var rs = stmt.executeQuery('select id, user, first_name, last_name from app_users');
  var document = SpreadsheetApp.getActive();
  var cell = document.getRange("a1");
  var row = 0;
  var rsMetaData= rs.getMetaData();
    for(var col=0;col < rsMetaData.getColumnCount(); col++){
        cell.offset(row, col).setValue(rsMetaData.getColumnName(col+1)); 
      } else {
        cell.offset(row, col).setValue(rs.getString(col+1));


The clear function is nothing more than

function clearSheet(){

This gives us a sheet with the menu:

And populated spreadsheet:

Since the code is run from google servers, they need to be able to communicate with your database server, with the IP address range of the google servers being defined on the JDBC guide:

Tuesday, 18 February 2014

Finding components with a particular build option

Say for example you have a generic build status of 'Exclude' - so you can temporarily disable components without touching the conditions (it doesn't really seem in the nature of build options, but still can be done). There will undoubtedly come a time where you need to locate all said components to either remove them from the application, or remove the build option so they are re-included.

First we can query the apex_dictionary view to find any views that contain the build option column:

select *
from apex_dictionary
where column_name = 'BUILD_OPTION'

As at ApEx 4.2, this gives us the following list:


So, now it's just a matter of writing a query against all of these views to locate the build option with the specified name. Common amongst all views are workspace, application id and build option. Then we can just add another field with relevant info specific to that view and the component type to help us narrow it down.

This will give us the following query:

variable cond varchar2(25)
exec :cond := 'Exclude';

, application_id
, info
, component_type
, build_option
from (

  select workspace, application_id, 'Display name: ' || display_name info, 'UI Type' component_type,  build_option
  from apex_appl_user_interfaces
  union all
  select workspace, application_id, 'Entry label: ' || entry_label || '; For page: ' || defined_for_page info, 'Breadcrumb entries' component_type, build_option
  from apex_application_bc_entries
  union all
  select workspace, application_id, 'Computation item: ' || computation_item info, 'Application computations' component_type, build_option
  from apex_application_computations
  union all
  select workspace, application_id, 'Item name: ' || item_name info, 'Application items' component_type, build_option
  from apex_application_items
  union all
  select  workspace, application_id, 'List name: ' || list_name info, 'Lists' component_type, build_option
  from apex_application_lists
  union all
  select workspace, application_id, 'Entry text: ' || entry_text || '; Parent entry:' || parent_entry_text info, 'List entries' component_type, build_option
  from apex_application_list_entries
  union all
  select workspace, application_id, 'LOV Name: ' || list_of_values_name || '; Display value: ' || display_value info, 'Static LOV entries' component_type, build_option
  from apex_application_lov_entries
  union all
  select workspace, application_id, 'Label: ' || icon_subtext info, 'Nav bar entries' component_type, build_option
  from apex_application_nav_bar
  union all
  select workspace, application_id, 'Page ID: ' || page_id info, 'Page' component_type, build_option
  from apex_application_pages
  union all
  select workspace, application_id, 'Page ID: ' || page_id || '; Branch point: ' || branch_point || '; Branch sequence: ' || process_sequence || '; Branch name: ' || branch_name info, 'Page branch' component_type, build_option
  from apex_application_page_branches
  union all
  select workspace, application_id, 'Page ID: ' || page_id || '; Item: ' || item_name || '; Computation point: ' || computation_point info, 'Page computation' component_type, build_option 
  from apex_application_page_comp
  union all
  select workspace, application_id, 'Page ID: ' || page_id || '; Dynamic action name: ' || dynamic_action_name info, 'Dynamic action' component_type, build_option
  from apex_application_page_da
  union all
  select workspace, application_id, 'Page ID: ' || page_id || '; Process name: ' || process_name || '; Process point: ' || process_point_code info, 'Page process' component_type, build_option
  from apex_application_page_proc
  union all
  select workspace, application_id, 'Page ID: ' || page_id || '; Region name: ' || region_name info, 'Page region' component_type, build_option
  from apex_application_page_regions
  union all
  select workspace, application_id, 'Page ID: ' || page_id || '; Region name: ' ||  region_name || '; Series name: ' || series_name info, 'Flash chart series' component_type , build_option
  from apex_application_page_flash5_S
  union all
  select workspace, application_id, 'Page ID: ' || page_id || '; Validation name: ' || validation_name info, 'Page validation' component_type, build_option
  from apex_application_page_val
  union all
  select workspace, application_id, 'Page ID: ' || page_id || '; Button name: ' ||  button_name info, 'Page button' component_type, build_option
  from apex_application_page_buttons
  union all
  select workspace, application_id, 'Page ID: ' || page_id || '; Item name' || item_name info, 'Page item' component_type, build_option
  from apex_application_page_items
  union all
  select workspace, application_id, 'Tab set name: ' || tab_set || '; Tab name: ' ||  tab_name info, 'Parent tabs' component_type, build_option
  from apex_application_parent_tabs
  union all
  select workspace, application_id, 'Process name: ' || process_name info, 'Application process' component_Type, build_option
  from apex_application_processes
  union all
  select workspace, application_id, 'Shortcut name: ' || shortcut_name info, 'Application shortcut' component_type, build_option
  from apex_application_shortcuts
  union all
  select workspace, application_id, 'Tab set: ' || tab_set || '; Tab name: ' || tab_name || '; Tab label: ' || tab_label info, 'Tabs' component_type , build_option
  from apex_application_tabs

Which you can then filter by build option, and voilĂ !

I've added this to a github project so it can be propely version controlled -