Geospatial Indexes

Redis Geospatial Indexes

Redis can store geographic coordinates (longitude/latitude) and efficiently query them by distance. Under the hood, geospatial data is stored in a sorted set using a geohash encoding, but Redis provides dedicated commands for spatial queries.

GEOADD — Store Locations

GEOADD restaurants -73.985428 40.748817 "Empire State Diner"
GEOADD restaurants -73.968285 40.763710 "Central Park Cafe"
GEOADD restaurants -74.006015 40.714270 "Wall Street Grill"

Format: GEOADD key longitude latitude member. You can add multiple members in a single command.

Note: Longitude comes before latitude in Redis commands (opposite of Google Maps' lat/lng order).

GEODIST — Distance Between Two Points

GEODIST restaurants "Empire State Diner" "Central Park Cafe" km

Supported units: m (meters, default), km, mi (miles), ft (feet).

GEOPOS — Get Coordinates

GEOPOS restaurants "Empire State Diner"

Returns the longitude and latitude of the stored member.

GEOSEARCH — Find Nearby Locations

GEOSEARCH restaurants FROMLONLAT -73.985 40.749 BYRADIUS 3 km ASC

Finds all members within 3 km of the given coordinates, sorted by distance (ASC = nearest first, DESC = farthest first).

You can also search from an existing member:

GEOSEARCH restaurants FROMMEMBER "Empire State Diner" BYRADIUS 5 km ASC

Real-World Use Cases

Ride-sharing (find nearby drivers):

GEOADD drivers -73.9857 40.7484 "driver:101"
GEOADD drivers -73.9712 40.7831 "driver:102"
GEOADD drivers -73.9903 40.7527 "driver:103"

GEOSEARCH drivers FROMLONLAT -73.985 40.750 BYRADIUS 2 km ASC COUNT 3

Store locator:

GEOADD stores -122.4194 37.7749 "SF Downtown"
GEOADD stores -122.4089 37.7837 "Union Square"
GEOADD stores -122.4332 37.7899 "Marina"

GEOSEARCH stores FROMLONLAT -122.42 37.78 BYRADIUS 5 km ASC

Delivery radius check:

GEOADD warehouse -73.985 40.748 "main-warehouse"
GEOADD customers -73.970 40.760 "customer:1"

GEODIST warehouse "main-warehouse" "customer:1" km
# If distance > 10km, charge delivery fee

How It Works Internally

Redis encodes each longitude/latitude pair into a 52-bit geohash and stores it as the score in a sorted set. This means:

• ZRANGE, ZCARD, and other sorted set commands also work on geo keys
• Memory usage is the same as a sorted set
• Range queries use the sorted set's O(log N) lookup

Your Task

Build a "find nearby coffee shops" feature:

1. GEOADD shops with these locations:
   • -73.985130 40.758896 "Times Square Bean"
   • -73.968285 40.763710 "Park Ave Coffee"
   • -73.993500 40.750600 "Hudson Yards Brew"
   • -74.006015 40.714270 "FiDi Espresso"
2. Find all shops within 4 km of longitude -73.985, latitude 40.755, sorted nearest first
3. Get the distance between "Times Square Bean" and "FiDi Espresso" in km