Redis Geospatial

Redis Geospatial commands store places as longitude and latitude pairs. They answer location questions such as "Which stores sit within 5 kilometers of me?" or "How far apart are two cities?" Delivery apps, ride services, and store locators rely on this feature.

The Map Pin Analogy

Picture a paper map covered with pins. Each pin carries a name and a position. A geospatial key works like that map. You drop pins with names, and Redis measures distances and finds the pins near any point you choose.

            North
              ▲
   ┌──────────┼─────────────────────┐
   │          │      ● Noida        │
   │   ● Delhi│                     │
   │          │                     │
   │  ● Gurgaon                     │
   └──────────┼─────────────────────┘
              ▼
            South

How Redis Stores Locations

Redis converts each coordinate pair into a single number and saves it in a Sorted Set. Every Sorted Set command still works on a geospatial key. You can list members with ZRANGE or delete one with ZREM.

GEOADD – Adding Places

Give longitude first and latitude second. Many beginners swap them by mistake.

127.0.0.1:6379> GEOADD cities 77.2090 28.6139 "Delhi" 77.3910 28.5355 "Noida" 77.0266 28.4595 "Gurgaon"
(integer) 3

Valid longitudes run from -180 to 180. Valid latitudes run from about -85.05 to 85.05.

GEOPOS – Reading Coordinates

127.0.0.1:6379> GEOPOS cities "Delhi"
1) 1) "77.20899999141693115"
   2) "28.61389934116891693"

The tiny differences from your input come from the conversion to a single number.

GEODIST – Measuring Distance

127.0.0.1:6379> GEODIST cities "Delhi" "Noida" km
"19.8"

Choose the unit with m, km, mi, or ft. The sample distance above is approximate. Redis returns a more precise decimal value.

GEOSEARCH – Finding Nearby Places

GEOSEARCH looks around a center point and returns members inside a radius or a box.

127.0.0.1:6379> GEOSEARCH cities FROMMEMBER "Delhi" BYRADIUS 22 km ASC WITHDIST
1) 1) "Delhi"
   2) "0.0000"
2) 1) "Noida"
   2) "19.8"

Search circle: radius 22 km around Delhi

        ┌───────────────────────┐
        │     ╭───────────╮     │
        │    ╱  ● Noida    ╲    │
        │   │   ● Delhi     │   │   ● Gurgaon (about 25 km) stays outside
        │    ╲             ╱    │
        │     ╰───────────╯     │
        └───────────────────────┘

Use FROMLONLAT in place of FROMMEMBER to search around raw coordinates, such as a phone's current position. Add COUNT 5 to limit the result list.

Older Commands

GEORADIUS and GEORADIUSBYMEMBER still run in current versions. Redis marks them as deprecated, so new projects should use GEOSEARCH.

Practical Example – Nearest Pharmacy Finder

Setup:
  GEOADD pharmacy 77.2167 28.6448 "Shop A" 77.2295 28.6129 "Shop B"

User opens the app at longitude 77.2190, latitude 28.6300:
  GEOSEARCH pharmacy FROMLONLAT 77.2190 28.6300 BYRADIUS 3 km ASC WITHDIST COUNT 3

App shows the three closest shops, nearest first, with distances.

One command replaces a heavy database query with map math.

Removing Places

127.0.0.1:6379> ZREM cities "Gurgaon"
(integer) 1

No GEODEL command exists, so ZREM does the job.

Accuracy and Limits

Redis computes distances with a spherical model of the Earth. The result stays within about 0.5 percent of the true distance, which suits delivery zones, store finders, and nearby-friend features. Surveying and aviation need exact geodesic math, so those fields should look elsewhere.

Geospatial keys grow like any Sorted Set. A key with millions of places still answers searches quickly, yet very large keys slow deletion and replication. Split data by region, such as pharmacy:north and pharmacy:south, when the list becomes huge.

Key Points

• GEOADD stores places as longitude then latitude.
• GEODIST returns the distance between two places.
• GEOSEARCH finds places inside a radius or box.
• Redis keeps geospatial data in a Sorted Set, so ZREM and ZRANGE work on it.