MongoDB provides the following geospatial index types that support geospatial queries.

• 2d indexes support queries that calculate geometries on a two-dimensional plane.
• 2dsphere indexes support queries that calculate geometries on an earth-like sphere.

In this article, I create a 2dsphere index.

Example Collection

Suppose we have a collection called bars with the following documents:

{
	"_id" : 1,
	"name" : "Boardwalk Social",
	"location" : {
		"type" : "Point",
		"coordinates" : [
			145.77675259719823,
			-16.919297718553366
		]
	}
}
{
	"_id" : 2,
	"name" : "The Downunder Bar",
	"location" : {
		"type" : "Point",
		"coordinates" : [
			145.77621640842125,
			-16.92107838010542
		]
	}
}
{
	"_id" : 3,
	"name" : "Riley",
	"location" : {
		"type" : "Point",
		"coordinates" : [
			145.7739955395154,
			-16.916028253292883
		]
	}
}
{
	"_id" : 4,
	"name" : "Salt House",
	"location" : {
		"type" : "Point",
		"coordinates" : [
			145.78148426655065,
			-16.91823513430776
		]
	}
}
{
	"_id" : 5,
	"name" : "Rattle n Hum",
	"location" : {
		"type" : "Point",
		"coordinates" : [
			145.77746095331537,
			-16.920051942529685
		]
	}
}

Each document has location information stored as a GeoJSON object.

A GeoJSON object has a field named type that specifies the GeoJSON object type and a field named coordinates that specifies the object’s coordinates.

Create the 2dsphere Index

Now let’s create the 2dsphere index.

db.bars.createIndex( { location : "2dsphere" } )

Output:

{
	"createdCollectionAutomatically" : false,
	"numIndexesBefore" : 1,
	"numIndexesAfter" : 2,
	"ok" : 1
}

The 2dsphere index has now been created.

We can now use the getIndexes() method to check our index:

db.bars.getIndexes()

Result:

[
	{
		"v" : 2,
		"key" : {
			"_id" : 1
		},
		"name" : "_id_"
	},
	{
		"v" : 2,
		"key" : {
			"location" : "2dsphere"
		},
		"name" : "location_2dsphere",
		"2dsphereIndexVersion" : 3
	}
]

We can see that the index was created as a 2dsphere index, using 2dsphereIndexVersion of 3, which is the default version for my current MongoDB installation (4.4).

Create a Compound 2dsphere Index

You can include 2dsphere index keys in compound indexes that are combined with non-geospatial index keys.

For example, we could have combined our location field with the name field to create a compound index.

Let’s drop the index and create a compound index:

db.bars.dropIndex("location_2dsphere")
db.bars.createIndex( { location : "2dsphere", name : 1 } )

Output:

{
	"createdCollectionAutomatically" : false,
	"numIndexesBefore" : 1,
	"numIndexesAfter" : 2,
	"ok" : 1
}

And check the index:

db.bars.getIndexes()

Result:

[
	{
		"v" : 2,
		"key" : {
			"_id" : 1
		},
		"name" : "_id_"
	},
	{
		"v" : 2,
		"key" : {
			"location" : "2dsphere",
			"name" : 1
		},
		"name" : "location_2dsphere_name_1",
		"2dsphereIndexVersion" : 3
	}
]

Compound 2dsphere indexes can reference multiple location and non-location fields. This is in contrast to compound 2d indexes, which are limited to referencing just one location field and one other field.

Changing the 2dsphereIndexVersion

You can change the 2dsphereIndexVersion by adding it as a field with the desired value when creating the index.

Example:

db.bars.createIndex( 
    { location : "2dsphere" },
    { "2dsphereIndexVersion" : 2 }
)

Output:

{
	"createdCollectionAutomatically" : false,
	"numIndexesBefore" : 1,
	"numIndexesAfter" : 2,
	"ok" : 1
}

Check the index:

db.bars.getIndexes()

Result:

[
	{
		"v" : 2,
		"key" : {
			"_id" : 1
		},
		"name" : "_id_"
	},
	{
		"v" : 2,
		"key" : {
			"location" : "2dsphere"
		},
		"name" : "location_2dsphere",
		"2dsphereIndexVersion" : 2
	}
]

This index has been created as a 2dsphere index version 2.

Version 2 is the default version of 2dsphere indexes created in MongoDB 2.6 and 3.0 series.

Version 3 is the default version of 2dsphere indexes created in MongoDB 3.2 and later (at the time of writing).

When I created the index without specifying the 2dsphereIndexVersion field, it created the index using version 3, because that’s the default version for my MongoDB version (4.4).