Question

A logistics application stores delivery vehicle positions in an Azure Cosmos DB for NoSQL container. Each document includes a GeoJSON Point in the 'location' property. The team runs frequent ST_DISTANCE queries to find all vehicles within a radius of a dispatch center, but these queries consume very high RU and perform slowly. The default indexing policy is in effect. What is the most effective change to reduce the RU cost of these geospatial queries?

Accepted Answer

Add a spatial index for the '/location/*' path with the Point data type in the indexing policy.

Answer

Add a composite index on '/location/coordinates/0' and '/location/coordinates/1'.

Answer

Switch the container's indexing mode from consistent to lazy so spatial lookups are cheaper.

Answer

Exclude the '/location/*' path from indexing to reduce write RU and speed up reads.