Quickie: Resizing a value prompt list

Another problem from a reader. Unlike checkbox or radio prompts, list prompts don’t expand and collapse to fit the number of options available to them.
This prompt is taking far too much room.

In the above image, we have a prompt that has too much space. Let’s shrink that down.
resizing lists

Now it shrinks to the number of rows displayed, or expands to a maximum of 15 items.

It’s actually a very easy script. The first step is to identify the prompt. We’ll use the old fW script (so this should work in every since 8.4).

var fW = (typeof getFormWarpRequest == "function" ? getFormWarpRequest() : document.forms["formWarpRequest"]);
if ( !fW || fW == undefined)
  fW = ( formWarpRequest_THIS_ ? formWarpRequest_THIS_ : formWarpRequest_NS_ );
var preFix = "";
if (fW.elements["cv.id"])
  preFix = fW.elements["cv.id"].value;
var nameSpace = "oCV" + preFix;

Next, we need to identify the prompt object:

var pr = fW._oLstChoicesPromptName;

The prompt itself is a standard SELECT element. We can set the size attribute to easily set the height.

This gives us:

if  (pr.options.length>15)

The following report XML is based on 10.2.1, but you can downgrade it easily by changing the schema number on the first line.
Resizing-Value-Prompt-List.txt (989 downloads)

Freezing headers in Cognos 10.2

Starting in Cognos 10.2, IBM released a way to freeze crosstab headers and rows. This method works in all major browsers, and doesn’t rely on dirty CSS hacks. To get it working, simply right-click on the list or crosstab, and select Freeze Headers. The report will even remember the state of the crosstab after the page refreshes (from a prompt, possibly).

right click action

While this method works well, there are many times where we’d want to have the lists or crosstab start off frozen; or maybe we want to give the user a button to freeze one or all of the crosstabs in one go. Somewhat surprisingly, the JavaScript is very easy to use.

First, let’s define the initial variables:

var paulScripts = {}
  , win=window['oCV'+'_THIS_'];

paulScripts.oCV = win.getRV().getCV();

Reports run from Report Studio need to use window[‘oCVRS’] while reports from the connection need window[‘oCV_NS_’]. Cognos will automatically replace _THIS_ to the correct fragment. Most of the Cognos functions we need are found inside win.getRV().getCV(), so we’re aliasing that into “paulScripts.oCV”.

The first function I’ll define is:

paulScripts.freezeContainer = function(objectName){
  var pFMngr = paulScripts.oCV.getPinFreezeManager();

All of the freezing functions are found inside getPinFreezeManager. The function getPinFreezeManager().freezeContainer() takes the object name (“Crosstab1”), and a Boolean to freeze/unfreeze the headers, and another Boolean to freeze/unfreeze the rows.

The freezing mechanism seems to have a small bug in the way it handles resizing. In my tests, it always resizes to a much smaller window than I need. The window.onResizeViewerEvent() function tricks the browser into thinking the window has been resized, and Cognos will then correctly recalculate the size of the frozen crosstab.

The unfreezing function is simple:

paulScripts.unfreezeContainer = function(objectName){
  var pFMngr = paulScripts.oCV.getPinFreezeManager();

These two functions will let us automatically freeze a crosstab as soon as the page loads. But say we want to toggle it on and off.

/* paulScripts.toggleContainer 
 * Paul Mendelson - 2014-03-25
 * This wil check if a container has frozen headers. If so, it unfreezes it, if not, it freezes the container. 
paulScripts.toggleAll = function(objectName) {
  var pFMngr = paulScripts.oCV.getPinFreezeManager();
  if(pFMngr.hasFrozenColumnHeadings(objectName)) {
    paulScripts.unfreezeContainer (objectName);
  else {
    paulScripts.freezeContainer (objectName);

The hasFrozenColumnHeadings function returns a Boolean, true or false, on the state of the column headers.

And finally let’s say that we want to freeze all the crosstabs on the page in one go.

/* paulScripts.toggleAll 
 * Paul Mendelson - 2014-03-25
 * in: type {string - crosstab or list}
 * This wil loop through every "type" on the page, and freezing or unfreezing, depending if the column header is frozen. 

paulScripts.toggleAll = function(type) {
    var xts = win._getContainers(type)
      , xtLen = xts.length
      , pFMngr = paulScripts.oCV.getPinFreezeManager();
      for(var i =0;i<xtLen;++i){
        var lid = pFMngr.removeNamespace(xts[i].getAttribute('lid'));
        if(i<xtLen-1 && lid==pFMngr.removeNamespace(xts[i+1].getAttribute('lid'))) continue; //when the panes are frozen the crosstab is split into four elements, all with the same lid. Without this hack, the xtab would toggle four times!
        if(pFMngr.hasFrozenColumnHeadings(lid)) {
        else {

That was a little bit more complex than before. We can use the _getContainers functions to get an array of all the lists or crosstabs on the page. The lid of the object is the name, plus the namespace. We can use the removeNamespace function to get it back to the name the freezeContainer functions expect.

The toggleAll function should only be used in a button. As Cognos stores the state of the object; if the headers are locked, they will remain locked after refreshing the page. Toggling them will cause them to unlock. Instead, it’s best to use a freezeAll function when loading the page:

paulScripts.freezeAll = function(type){
  var xts = win._getContainers(type)
  , xtLen = xts.length
  , pFMngr = paulScripts.oCV.getPinFreezeManager();
  for(var i =0;i<xtLen;++i){
    var lid = pFMngr.removeNamespace(xts[i].getAttribute('lid'));
    if(pFMngr.hasFrozenColumnHeadings(lid)) continue; 
    paulScripts.freezeContainer (lid);

Once everything is working, we can see how it works.
Freezing panes

The example report in using 10.2.1, against the sales and marketing cube.
Freezing Panes report XML (1877 downloads)

Passing parameters without refreshing the page

Recently a reader had an interesting problem. Her client needed a prompt page from which they could open different reports. As it stands now, the client needed to select their prompts, hit the finish button, and then click on the links. She wanted a way to skip the second step – select the prompts, and have those values pass in the drill through.

With the prompt API, and a bit of JS, this is incredibly easy.

The first step is to identify the target report. You can get the report search path through the properties. Click on the “View the search path, ID and URL Location” link, and a window will pop up with the path and other details.
1. Get search path

Notice the double quotes? That can potentially cause trouble with the JavaScript, so I recommend using a URL Encoder: /content/folder[@name=’JavaScript Examples’]/folder[@name=’Drill Target’]/report[@name=’Target’] becomes %2Fcontent%2Ffolder%5B%40name%3D%27JavaScript%20Examples%27%5D%2Ffolder%5B%40name%3D%27Drill%20Target%27%5D%2Freport%5B%40name%3D%27Target%27%5D

The example I’m posting will be using the URL Encoded version.

The next step will be the JavaScript:


  var paulScripts = {}
    , oCR = cognos.Report.getReport("_THIS_")
    , gateway = window['oCV_THIS_'].getRV().getCV().sGateway
    , report1 = decodeURIComponent("%2Fcontent%2Ffolder%5B%40name%3D%27JavaScript%20Examples%27%5D%2Ffolder%5B%40name%3D%27Drill%20Target%27%5D%2Freport%5B%40name%3D%27Target%27%5D%0A%0A");

  /* function paulScripts.URLEncode - Paul Mendelson = 2013-10-21
   * Based on the JSONEncode function, this will return the use/display values of
   * a prompt in a format that can be based via POST or GET.
   * usage: 'p_Parameter1='+paulScripts.URLEncode(paulScripts.getControl('MyPrompt'))
  paulScripts.URLEncode = function(promptControl){
    var urlData = '<selectChoices>'
      , aPromptValues = promptControl.getValues()
      , nonRange = []
      , range =[];

    if (aPromptValues.length == 0) {return false}
    for (var j=0; j< aPromptValues.length; j++) {
      var promptValue =  aPromptValues[j];
      if (promptValue.use) {// Non Range value
        nonRange.push("<selectOption useValue='"+promptValue.use + "' displayValue='"+ promptValue.display + "'/>");
      else { // Range value   
        var rangeStart = promptValue.start
          , rangeEnd = promptValue.end
          , start, end, startDisp, endDisp;
        if (rangeStart && rangeEnd) { //has both Start and End (bounded)
          rangeStart.display?startDisp = " displayValue='"+ rangeStart.display+"'":"";
          rangeEnd.display?endDisp = " displayValue='"+ rangeEnd.display+"'":"";
          range.push("<selectBoundRange><start useValue='"+rangeStart.use +"'" + startDisp +"/><end useValue='"+rangeEnd.use +"'" + endDisp +"/></selectBoundRange>")
        else if (rangeStart && !rangeEnd) {//unboundedEndRange
          rangeStart.display?startDisp = " displayValue='"+ rangeStart.display+"'":"";
          range.push("<selectUnboundedEndRange><start useValue='"+rangeStart.use +"'" + startDisp +"/></selectUnboundedEndRange>");
        else if (!rangeStart && rangeEnd) {
          rangeEnd.display?endDisp = " displayValue='"+ rangeEnd.display+"'":"";
          range.push("<selectUnboundedStartRange><end useValue='"+rangeEnd.use +"'" + endDisp +"/></selectUnboundedStartRange>");
        else {
          alert ("Range not set.");
        } // end if
      } // end if
    } // end for
    if(nonRange.length>0) urlData += nonRange.join();
    if(range.length>0) urlData += range.join();
    return urlData.replace(/&/g,'&amp;');

  /* function paulScripts.getPrompts
   * Paul Mendelson - 2013-10-21
   * Retrieves all the prompts on the page and returns the parameter values
  paulScripts.getPrompts = function(){
  //find all of the controls and stick them in aPromptControls.
    var aPromptControls = oCR.prompt.getControls( ),
    prompts = [];

  //loop through the controls - finding the name and values and sticking them into the names array
      if(aPromptControls[i].getValues().length == 0) continue;
      prompts.push(paulScripts.URLEncode (aPromptControls[i]));
  return  prompts;

paulScripts.runReport = function(reportName,reportPath,gateway, prompts){
  var basicInfo = [
          reportName, "toolbar=no,height=200,width=400"
        , 'ui.gateway' , gateway
        , 'ui.tool' , 'CognosViewer'
        , 'ui.action' , 'run'
        , 'ui.object' , reportPath
        , "run.prompt" ,"false"
        , "cv.toolbar" ,"false"
        , "cv.header" ,"false"
        , "run.prompt", "false"

  return cognosLaunchInWindow.apply(this,basicInfo);


Let’s go through the functions.

var paulScripts = {}
, oCR = cognos.Report.getReport(“_THIS_”)
, gateway = window[‘oCV_THIS_’].getRV().getCV().sGateway
, report1 = decodeURIComponent(“%2Fcontent%2Ffolder%5B%40name%3D%27JavaScript%20Examples%27%5D%2Ffolder%5B%40name%3D%27Drill%20Target%27%5D%2Freport%5B%40name%3D%27Target%27%5D%0A%0A”);

We’re defining the variables on the page. Notice that I’m hardcoding the report search path into a variable here. There’s actually no reason to do that, we could put it directly into the function, or put it into another prompt, or even insert it into a table and build a list with each row calling another report.

The gateway variable is pulling the gateway from the report you’re running. Another gateway could be entered here, if needed.

The paulScripts.URLEncode will take a prompt control, process the selected values, and spit out a string that Cognos can use to populate parameters.

paulScripts.getPrompts will look for every prompt on the page, and create an array of the parameter names and the selected strings. This makes updating the report simple – you won’t need to modify and JavaScript when adding or removing prompts.

paulScripts.runReport is the last function. It uses the cognosLaunchInWindow function to run a report in the specified window, using the specified search path, on the specified gateway, with the results of the paulScripts.getPrompts function.

To use it, all you need to do is creating an onclick event like this:

<input type="button" onclick="paulScripts.runReport('_blank',report1,gateway, paulScripts.getPrompts())" value="Run Report 1"/>

And finally, we can test it:

This report was built on 10.2.1, but should work on all versions 10.2 and up.
Passing Parameters Without Refresh (4766 downloads)