more than one 2dsphere index, not sure which to run geoNear

Question

I am using $geoNear and near inside the aggregate in MongoDB. I have hosted my MongoDB database into mlabs. And everything working fine on my local, But don't know why when I deploy my app to live and I am getting the below error:

"geoNear command failed: { ok: 0.0, errmsg: \"more than one 2dsphere index, not sure which to run geoNear on

Below is the code which I have used:

Shops.aggregate([
  {
     $geoNear: {
         near: { 
            type: "Point",
            coordinates: coordinates
         },
         distanceField: "dist.calculated",
         maxDistance: 80467,
         spherical: true
     }
  }
])
.then((products)=>{
      res.json(products);
})

Can anyone please help me with the same?

Answer 1

As the error message indicated, this is because you have more than one 2dsphere indexes, so $geoNear doesn't know which one to use.

In this situation, you can either:

• drop the second geo index, or
• use the key parameter as mentioned in the $geoNear documentation:

If your collection has multiple 2d and/or multiple 2dsphere indexes, you must use the key option to specify the indexed field path to use. Specify Which Geospatial Index to Use provides a full example.

The error is mentioned in the docs as well:

If there is more than one 2d index or more than one 2dsphere index and you do not specify a key, MongoDB will return an error.

You can use db.collection.getIndexes() to list all indexes defined on the collection.

Here's an example of using the key parameter:

> db.test.insert([
  {_id:0, loc1:{type:'Point',coordinates:[1,1]}, loc2:{type:'Point',coordinates:[2,2]}},
  {_id:1, loc1:{type:'Point',coordinates:[2,2]}, loc2:{type:'Point',coordinates:[1,1]}}
])

Then I create two 2dsphere indexes:

> db.test.createIndex({loc1:'2dsphere'})
> db.test.createIndex({loc2:'2dsphere'})

Running $geoNear without specifying key will output the error:

> db.test.aggregate({$geoNear:{near:{type:'Point',coordinates:[0,0]},distanceField:'d'}})
...
  "errmsg": "more than one 2dsphere index, not sure which to run geoNear on",
...

Using key: loc1 will sort the result according to the loc1 index (_id: 0 comes before _id: 1):

> db.test.aggregate(
    {$geoNear: {
        near: {type: 'Point',coordinates: [0,0]},
        distanceField: 'd',
        key: 'loc1'}})
{ "_id": 0, "loc1": { "type": "Point", "coordinates": [ 1, 1 ] }, "loc2": { "type": "Point", "coordinates": [ 2, 2 ] }, "d": 157424.6238723255 }
{ "_id": 1, "loc1": { "type": "Point", "coordinates": [ 2, 2 ] }, "loc2": { "type": "Point", "coordinates": [ 1, 1 ] }, "d": 314825.2636028646 }

And, using key: loc2 will sort the result according to the loc2 index (_id: 1 comes before _id: 0):

> db.test.aggregate(
    {$geoNear: {
        near: {type: 'Point',coordinates: [0,0]},
        distanceField: 'd',
        key: 'loc2'}})
{ "_id": 1, "loc1": { "type": "Point", "coordinates": [ 2, 2 ] }, "loc2": { "type": "Point", "coordinates": [ 1, 1 ] }, "d": 157424.6238723255 }
{ "_id": 0, "loc1": { "type": "Point", "coordinates": [ 1, 1 ] }, "loc2": { "type": "Point", "coordinates": [ 2, 2 ] }, "d": 314825.2636028646 }