This snippet enables to iterate over a Firebase reference in a non-blocking way. If you need to iterate over a large reference, a child_added query may block your Firebase as it will query the whole data before iteration. With this snippet, children are retrieved one by one, making it slower / safer.

Firebase Lazy-Safe Iterator

// By Clément Wehrung
  
function Iterator(queueRef, processingCallback) {
  this.queueRef = queueRef;
  this.processingCallback = processingCallback;
  this.processed = {};
  this.processNext();
}

Iterator.prototype.processNext = function() {

  var priority = this.currentItem ? this.currentItem.getPriority() : undefined,
      name = this.currentItem ? this.currentItem.name() : undefined;
      
  this.queueRef.startAt(priority, name).limit(2).once("value", function(listSnap) {

    listSnap.forEach(function(snap) {
      
      var name = snap.name();
          
      if (this.processed[name]) return;
    
      this.processingCallback( snap, function() {
  
        if (this.currentItem) {
          setTimeout(this.processNext.bind(this), 0);
        }
        
      }.bind(this) );

      this.currentItem = snap;    
      this.processed[ name ] = true;
           
    }.bind(this));
     
  }, this);
  
}

module.exports = Iterator;

Thanks for the helpful snippet! May I ask why this line is needed: if (this.processed[name]) return;

Why might children be processed twice with child_added?

The issue is that child_added will also be triggered with the last item. This line at least makes it obvious. I have in the meantime found one major issue with this version of the code, as the .on('child_added' are never switched off. Pushing a safer version as this one may lead to firebase using up to 100% CPU