Overview

EVAL and EVALSHA are Redis commands used to execute Lua scripts within the Redis server. They differ primarily in how they handle script loading and execution:

EVAL: This command takes the Lua script as an argument. Redis parses and hashes the script every time EVAL is called. This adds overhead, especially for frequently executed scripts.
EVALSHA: This command takes the SHA1 hash of the Lua script as an argument. Redis retrieves the script from its internal cache using this hash. If the script is found, it’s executed directly, bypassing the parsing and hashing steps. This results in significantly faster execution times for frequently used scripts.

Key Differences Summarized

Feature	`EVAL`	`EVALSHA`
Script Input	Lua script as a string	SHA1 hash of the Lua script
Script Parsing	Parses and hashes the script each time	Retrieves from cache; no parsing
Performance	Slower, especially for frequent use	Faster for frequently used scripts
Script Cache	No reliance on script cache	Relies on script cache; error if miss
Error Handling	No specific error for cache miss	Returns error if script not in cache

When to Use Which

EVAL: Use for testing, debugging, or scripts that are executed infrequently. It’s simpler to use since you don’t need to manage script hashes.
EVALSHA: Use for frequently executed scripts to optimize performance. This requires a prior step to load the script using SCRIPT LOAD to obtain its SHA1 hash.

Example

Let’s say you have a Lua script to increment a counter:

redis.call('INCR', KEYS[1])

Using EVAL:

EVAL "redis.call('INCR', KEYS[1])" 1 mykey

Using EVALSHA:

a. Load the script and get the SHA1 hash:

SCRIPT LOAD "redis.call('INCR', KEYS[1])"

Redis computes the script’s SHA1 hash and stores both the script and its hash in an internal cache.

This returns the SHA1 hash (e.g., a1b2c3d4e5f6...).

b. Execute using the hash:

EVALSHA a1b2c3d4e5f6... 1 mykey

Remember to replace a1b2c3d4e5f6... with the actual SHA1 hash. EVALSHA will be significantly faster if this script is executed repeatedly. However, if the script is not in the cache, EVALSHA will fail. Therefore, robust error handling is crucial in production environments. A common strategy is to use EVALSHA first and fall back to EVAL if it fails.

Redis