Dragonfly Cloud announces new enterprise security features - learn more

Golang: Deleting Redis Keys by Prefix (Detailed Guide w/ Code Examples)

Use Case(s)

In Golang applications that utilize Redis as a key-value store, it might be necessary to delete keys based on a certain prefix. This can be used for data cleanup processes or when related keys need to be removed at once.

Code Examples

The following example shows how to use the Scan function with a specified prefix and then remove the matching keys. Please install the go-redis/redis package before running this code.

package main import ( "github.com/go-redis/redis/v8" "context" "fmt" ) func main() { rdb := redis.NewClient(&redis.Options{ Addr: "localhost:6379", Password: "", // no password set DB: 0, // use default DB }) ctx := context.Background() var cursor uint64 var n int for { var keys []string var err error keys, cursor, err = rdb.Scan(ctx, cursor, "myprefix*", 10).Result() if err != nil { panic(err) } for _, key := range keys { fmt.Println("deleting key: ", key) rdb.Del(ctx, key) } if cursor == 0 { break } } fmt.Printf("Deleted %d keys\n", n) }

This code connects to a local Redis instance, then it scans keys that match the prefix myprefix. Matching keys are then deleted using Del function.

Best Practices

  • Make sure to limit the number of keys per Scan operation to avoid blocking the server for long periods.
  • Always handle the errors returned by the Scan and Del functions.

Common Mistakes

  • A common mistake is to use the KEYS command followed by DEL in production environments. KEYS can block the server when used against large databases, prefer SCAN instead.

FAQs

Q: Can I delete keys with multiple different prefixes at once? A: No, you need to run separate Scan and Del operations for each prefix.

Q: What happens if there are no keys matching the prefix? A: The Scan function will simply return an empty list, and no keys will be deleted.

Was this content helpful?

Switch & save up to 80% 

Dragonfly is fully compatible with the Redis ecosystem and requires no code changes to implement. Instantly experience up to a 25X boost in performance and 80% reduction in cost