Skip to main content

Push changed rows to an Interactive Grid

For pushing changes from the database to the end user, the regular solution is using websockets. A change in a record is detected - using a trigger or using the CQN (Change Query Notification) feature - and a notification is send to a websocket server. That websocket server broadcasts the notification over a channel to all browsers that are tuned in to that websocket channel. Then the browser reacts to that notification, usually showing an alert or refreshing a report. This trick is described on multiple sites, just Google for "oracle apex websockets" or similar.

So back in the old days, we used that notification in the browser to refresh the (interactive) report. But along comes the Interactive Grid (IG). While he full-refresh mechanism still works for IG, an IG has also the option to refresh just one row. 
So wouldn't it be awesome that just the changed row(s) get refreshed upon a change in the database, instead of the whole report? Can we do it ... yes we can!

First in our "Execute when Page Loads" property we start listening to the websocket server when the page is loaded using:

window.addEventListener("load", init, false);

In the "Function and Global Variable Declaration" we add the actual code and the init function:

//The websocket server is set using an Application Item 
var wsUri = "&WSSERVER.";

function init() {
  websocket = new WebSocket(wsUri); 
  websocket.onopen = function(evt) { onOpen(evt) };
  websocket.onclose = function(evt) { onClose(evt) };
  websocket.onmessage = function(evt) { onMessage(evt) };
  websocket.onerror = function(evt) { onError(evt) };
}

// This is where the real magic happens
// onMessage is called when a message is received from the websocket server 
function onMessage(msg) {
  // msg is in JSON format, containing the ROWID of the changed record
  var rowid, 
      record, 
      records = [],
      json = JSON.parse( msg.data ),
      l = json.changes.length;  
  // "myGrid" is the Static ID of the IG region
  var myGrid = apex.region("myGrid").widget().interactiveGrid("getViews").grid;
  for ( var i = 0; i < l; i++){   
    rowid = json.changes[i].rowid ;
    // loop through all the rows, check if the rowid is present
    // forEach callback params : record, index, key
    myGrid.model.forEach( function(r, i, k){
      if ( $.inArray( rowid, r ) > -1 ) {
        record = myGrid.model.getRecord(k);
        if ( myGrid.model.getRecordMetadata(k).updated ){
          apex.message.alert("Record " + (i+1) + " is changed by another user and is requeried");
        }
        records.push( record );
      }
    });
  }
  // fetch the changed records
  if ( records.length > 0 ) {
    myGrid.model.fetchRecords(records);  
  }
}

// Addtional helper functions below
$(window).on("beforeunload", function(){
  websocket.close();
});  

function onOpen(evt) {
 apex.debug("Websocket opened on &WSSERVER.");
}

function onClose(evt) {
  apex.debug("Websocket closed");
}

function onError(evt) {
  apex.debug("Websocket Error");
  apex.debug(evt);
}

The key of the code above is that the records that are changed and are found on the current page are added to an array of records. And that array us used to fetch the records. If you inspect the network traffic that is going on, you can verify that only the changed records (in the screen shot below, three changed records) are fetched and shown on the screen.

The whole process works seamless for the user. If he made a change to a record that didn't get updated by someone else (or another process) than he can keep working on that, if he did make a change - and thus there is a conflict - he will get a notification. In our example the process didn't even show a progress spinner, but maybe that's because our queries are fast (it is not on EMP, but based on a view that references several tables and returns over 40,000 rows).

[Edit Jul 4 : I changed the code in a way that it is dependent on the DOM anymore, but only uses the model to determine whether there is a record to update]]

Comments

Popular posts from this blog

Filtering in the APEX Interactive Grid

Remember Oracle Forms? One of the nice features of Forms was the use of GLOBAL items. More or less comparable to Application Items in APEX. These GLOBALS where often used to pre-query data. For example you queried Employee 200 in Form A, then opened Form B and on opening that Form the Employee field is filled with that (GLOBAL) value of 200 and the query was executed. So without additional keys strokes or entering data, when switching to another Form a user would immediately see the data in the same context. And they loved that. In APEX you can create a similar experience using Application Items (or an Item on the Global Page) for Classic Reports (by setting a Default Value to a Search Item) and Interactive Reports (using the  APEX_IR.ADD_FILTER  procedure). But what about the Interactive Grid? There is no APEX_IG package ... so the first thing we have to figure out is how can we set a filter programmatically? Start with creating an Interactive Grid based upon the good old Employ

apex_application.g_f0x array processing in Oracle 12

If you created your own "updatable reports" or your custom version of tabular forms in Oracle Application Express, you'll end up with a query that looks similar to this one: then you disable the " Escape special characters " property and the result is an updatable multirecord form. That was easy, right? But now we need to process the changes in the Ename column when the form is submitted, but only if the checkbox is checked. All the columns are submitted as separated arrays, named apex_application.g_f0x - where the "x" is the value of the "p_idx" parameter you specified in the apex_item calls. So we have apex_application.g_f01, g_f02 and g_f03. But then you discover APEX has the oddity that the "checkbox" array only contains values for the checked rows. Thus if you just check "Jones", the length of g_f02 is 1 and it contains only the empno of Jones - while the other two arrays will contain all (14) rows. So for

Stop using validations for checking constraints !

 If you run your APEX application - like a Form based on the EMP table - and test if you can change the value of Department to something else then the standard values of 10, 20, 30 or 40, you'll get a nice error message like this: But it isn't really nice, is it? So what do a lot of developers do? They create a validation (just) in order to show a nicer, better worded, error message like "This is not a valid department".  And what you then just did is writing code twice : Once in the database as a (foreign key) check constraint and once as a sql statement in your validation. And we all know : writing code twice is usually not a good idea - and executing the same query twice is not enhancing your performance! So how can we transform that ugly error message into something nice? By combining two APEX features: the Error Handling Function and the Text Messages! Start with copying the example of an Error Handling Function from the APEX documentation. Create this function