Add hasLayer() to the markercluster object

[grantlucas]

When adding new markers to a map, it'd be great to be able to easily check if it already has that layer. It's pretty simple to do without this function but it would be a nice added feature.

[Adventsparky]

+1, This would be nice.

[danzel]

How about if calling addLayer when the layer was already added it didn't re-add the same marker?

[grantlucas]

Hm. I can see how that would be good but I can also see cases where you just want to check if that layer is there. I would be more for the hasLayer method to keep it similar to what the normal map object has. I also wonder if there may be some reason to add a duplicate layer.

[danzel]

Yep fair enough. Just a bit hard in the code ATM as we don't keep a list of all markers around to easily check against, so we'll need to traverse the tree.

I don't think you should be adding the same marker to the map twice. Currently if you do that with the clusterer it clusters with it self and results in all sorts of weirdness.

[grantlucas]

Ah true true about duplicate markers.

Right now, my duplicate detection is working like so. (pretty much ripped it from hasLayer on the map object).

for(var id in me.markers)
{   
    var layerId = L.Util.stamp(me.markers[id].marker);
    if(!me.markersCluster._layers.hasOwnProperty(layerId))
        me.markersCluster.addLayer(me.markers[id].marker);
}

Seems to be working alright so far.

[danzel]

I think that will only work if the given marker is currently visible on the map.
If it is inside a cluster at the visible zoom level then it will not be in _layers so you'll add it again.

[grantlucas]

Ah that makes sense. I just checked over my code and in a host of changes in the past week, I'm reseting the entire cluster markering since I'm loading markers through AJAX while moving around the map so my check doesn't even make sense anyways. I had added it at one point when I wasn't resetting the layer.

[Adventsparky]

Any plans to incorporate the check if a layer exists when adding as mentioned? That alone would be nice. the hasLayer() would be great too, but if the check when adding is easier, that's good too.

[danzel]

Yep, might be looking at this later today, we'll see how my day unfolds.

Out of personal interest, what sort of projects are you guys using the markercluster on?
If it is something private feel free to send me an email (its on my github profile) or if it is something public it would be nice to be able to link to it to show that people are actually getting good use of the project :-)
Thanks!

[danzel]

Alrighty! I've added this and also made addLayer check if the layer is already in the cluster before adding it.
You can disable the test for a small performance boost if you know you'll only ever add markers that aren't in the clusterer.
The test only runs once the MarkerClusterGroup has been added to the map, so initial load performance will still be the same.
Enjoy!

[cvisto]

Hello. In my application sometime I need to readd markers that were removed from clusterGroup. hasLayer(marker) method returns "true" if marker was removed from cluster. I think it's because at that moment marker.__parent._group === clusterGroup. To my mind it would be better if hasLayer returns correct answer for remove markers (doesn't metter were they removed by removeLayer[s] or clearLayers).

What do you think about that?

[cvisto]

Now there are 4 ways of adding layers to clusterGroup:

1. addLayer() when cluster IS NOT on the map
2. addLayers() when cluster IS NOT on the map
3. addLayer() when cluster IS on the map
4. addLayers() when cluster IS on the map

Now only one of them (3rd) checks for duplicates. Maybe you can make all of them do that checking. And add some MarkerClusterGroup option (for example, "checkDuplicates: true") to allow developer to decide if he wants that checking (with understanding that it will decresase productivity) or not (in order to increase performance)