Getting the Select2-to-Tree working in XPages

In a modernization project (increase browser compatibility) I needed a more simple solution for a custom value picker. The less design elements the merrier 🙂

Another highly desired feature was to have a search function within the picker because there are a lot of values to choose from.

So my initial though was to present the options in a listbox and wrap the select2 plugin around it. Well it turned out the customer found the categorization of options a need to have, including collapse and expand options (who has not grown up with twisties?).

Next attempt was to implement the Select2-to-tree plugin since it presents a hierarchy of links in a select2 object

Looks familiar like a Notes view right? Luckily my Notes view only got one level of categorization and my skills with transforming Notes view data into other objects came in handy. So in a short time I have my data providing REST service setup .

Here is the initiation script for the plugin:

Then I tried to send the selected value to the server to check the value but I was not able to. The XPage did not send a message but the page just frooze.

So to avoid to spend a lot of time on finding and fixing the bug I decided to run the alternative route and use SELECT element instead of the xp:listBox and write the value from csjs to the server via a RPC service (hence the change listener)


The RPC service has nothing special, just setting a viewScope variable:

So that is basically it. With a minimum of design elements / coding I got my Notes view data in a treeview presented in a select2 component:

One item to be removed from my back-log. Happy development 🙂

Data for Bootstrap Treeview from a Notes View?

In a modernization project (increase of browser compatibility) I needed to find a solution for a list of links, categorized and sorted like a Notes View.

Since I already use Bootstrap as CSS framework I decided to check the following Bootstrap Treeview project: .  I was satisfied with the following example(s):


The initialisation is pretty simple:


Now I just needed to set up a REST service to provide me the data. The service is setup on an XPage and bounded to a Java class:


The Java class I will reuse for providing multiple data streams so it detects the different method parameters:


As mentioned the data for the list comes from a Notes View. As you all know views entries have different characteristics which I have to bear in mind:

  • Documents can land anywhere in the view since it is categorized on a text field where the multiple levels can set individually eg Europe\\Sweden\\Stockholm or Afrika\\Johannesburg.
  • The view also holds links for other menus organized under a category so I decided to us a viewnavigator and start collecting data from a specific category.
  • For each entry in the view I have to check what type it is: category, document or total. The last one is not of my interest so I have to skip if that option occurs.
  • The columnindentlevel tells me all about where I am in the view in comparison with the columnindentlevel of the previous entry. Here are the scenarios:

// case 1: current indent level < columnindentlevel
// -> new category, propobably start situation
// case 2: current indent level = column indentlevel
// -> new category (sibling) but close the previous
// one first (just one level)
// case 3: current indent level > column indent
// level -> new category but closes the previous
// one(s) first. how many depends on difference curr
// and column level

Ofcourse categories behave different that documents.

So here is the code:


Some notes:

  • I have set the header of the REST service to text/html and the plugin needs a JavaScript object. Therefor capture my data response in an eval() method.
  • A target attribute is not provided by the plugin, so I add one myself. Categories have the # as href so based on that info I include a target attribute or not. I do this via a function:
  • If the treeview is collapsed there is no anchor element for underlying list-items so you cannot add the target for all links.
  • Also when collapsing and expanding a category the added target attributes are gone. So for opening of every category you need to re-apply the href attribute for underlying anchors.
  • A category can be opened via a ‘twistie’ image or the text link so we need to register an action on these onclick events

The result is a nice looking ‘Notes View data-driven’ treeview with Bootstrap styling:


Happy coding 🙂

PS. I noticed I have a lot of unused local variables in my code, you may clean that up 🙂

Building a search function with DataTables plugin (VI)


Often business analists who work with the data in tables eat & sleep in their spreadsheet world so a common asked question is: can we export the data? “Ofcourse you can!” with the datatable component.


Below is a short summary of the changed/added items.


In order to add the buttons you specify in the datatable configuration which buttons you would like to have e.g.:

dom: ‘Bfrtip’,
buttons: [
extend: ‘pdfHtml5’,
orientation: ‘landscape’,
pageSize: ‘LEGAL’,
title: ‘Person Details’


Required files

Which files are required for the buttons feature depend on which buttons you would like to display. In the download builder you can specify for which options you would like to go.

AMD fix

An important thing to notice is that most JS files use AMD loading. There have been posted multiple questions and answers how to fix AMD loading with XPages. In my repository I have just removed the amd checking.

Files overview

Probably the files you would like to use will be listed in the Theme design element:

<!– adding buttons –>



As a result you will see the buttons displayed above the table. If you apply filtering to the table the buttons will pick up the provided terms.


A little less coding than an X-agent to create a PDF or spreadsheet 🙂

A downsize might be that users are not aware of the amount of data that might reside in your datatable (the only x-number of entries are displayed). Pressing the PDF button with 40K of documents in your notes view takes a bit of time…

Navigation list from Notes view in Twitter Bootstrap (with XPages)


In my previous post I demonstrated how you create breadcrumbs from Notes view in Twitter Bootstrap. A typical Notes application contains an outline (I assume you don’t the Navigator design element anymore). In this post I demonstrate how to create something similar in Twitter Bootstrap, a Nav List.

navlist example

Nav Lists

A Nav list is a simple and easy way to build groups of nav links with optional headers. They’re best used in sidebars like the Finder in OS X. You can nest nav-lists and then they come quit close to categorized or hierarchical structured documents (e.g. parent & response).


As  a start database I used Mark Leusink’s Bootstrap4Xpages demo site which you can download here.

As source I am also using a similar view as I used in my previous post. Here is how it looks like:

notesview clean

I found an example code for nesting lists on Fiddle. Note that you can extend nav-list with mouse over behavior but I fear for mobile phones and tablets this will be a crime.

I display the results via a computed field that looks as followed:

<xp:text escape=”false” id=”computedField2″ value=”#{javascript:getNavList()}”></xp:text>

The function getNavList() looks as followed:

function getNavList(){
var nav:NotesViewNavigator = database.getView(“$v-treeview-clean”).createViewNav();
var entry:NotesViewEntry = nav.getFirst();
if (entry !=null){
var countLevel:Integer = 0;
var curLevel:Integer;
var list = “<div class=\”well sidebar-nav\”><ul class=\”nav nav-list\”>”;
while (entry !=null){
var edoc:NotesDocument = entry.getDocument();
entryValue = entry.getColumnValues().elementAt(1).toString();
var col:NotesDocumentCollection = edoc.getResponses();
curLevel = entry.getColumnIndentLevel();
if (col.getCount()>0){
//prep for new entry with response(s)
var difLevel = countLevel – curLevel ;
var closure = “”
for (var i=0; i<(difLevel); i++){
closure = closure + “</ul></li>”
list = list + closure;
list = list + “<li>”;
list = list + “<a href=\”#\” data-toggle=\”collapse\” data-target=\”#” + entryValue + “\”>” + entryValue + “</a>”;
list = list + “<ul id=\”” + entryValue + “\” class=\”collapse in\”>”;
//increase counter
countLevel = curLevel + 1;
list = list + “<li><a href=\”#\”>” + entryValue + “</a></li>”;
else if(curLevel<countLevel){
var difLevel = countLevel – curLevel ;
var closure = “”
for (var i=0; i<(difLevel); i++){
closure = closure + “</ul></li>”
list = list + closure
countLevel = curLevel;
else {
var tmpentry:NotesViewEntry = nav.getNext(entry);
entry = tmpentry;
//final closure, last entry could be response doc
var closure = “”
for (var i=1; i<(countLevel); i++){
closure = closure + “</ul></li>”
list = list + closure
//closure nav nav-list
list = list + “</ul>”;
//closure well sidebar-nav
list = list + “</div>”;
return list;
return “no documents found”;

In order to make it look nice embed the computed field in div element and give it a span class.

Below you see the final result:

result nav list



You notice directly an obvious problem: the long title references which are typical for Notes documents.


Twitter Bootstrap offer other alternatives which might be more suitable for transforming Notes views. And example is the dropdown. This you can extend with the dropdown JavaScript plugin.

Point for improvement

A class I did not use for the LI element is nav-header. This gives a nice outstanding header, which are great for categorized views.

A class I did not use for the LI element is active which highlights a list item. When using the nav-list on documents opened via an XPage I guess it is best practice to add the class name when the document is ready on the client.

The HREF attribute is not computed but this can be extracted from the viewentry.

Performance. The view above contains very little documents. I still have to test it with larger databases. Some examples where I want to use the nav-list contain more than 15000 documents in the view. I wonder how much I can use sessionScope or place the HTML in a profile document. Any suggestions here are welcome.

Tools used

Beside DDE is used heavily NotePad++ to check the generated HTML.

Job Wanted

Looking for a creative brain? Choose me!