DuckDB Redis Client Extension

This extension provides Redis client functionality for DuckDB, allowing you to interact with a Redis server directly from SQL queries.

Experimental: USE AT YOUR OWN RISK!

Features

Currently supported Redis operations:

• String operations: GET, SET, MGET
• Hash operations: HGET, HSET, HGETALL, HSCAN, HSCAN_OVER_SCAN
• List operations: LPUSH, LRANGE, LRANGE_TABLE
• Key operations: DEL, EXISTS, TYPE, SCAN, KEYS
• Batch and discovery operations: SCAN, HSCAN_OVER_SCAN, KEYS

Quick Reference: Available Functions

Function	Type	Description
redis_get(key, secret)	Scalar	Get value of a string key
redis_set(key, value, secret)	Scalar	Set value of a string key
redis_mget(keys_csv, secret)	Scalar	Get values for multiple keys (comma-separated)
redis_hget(key, field, secret)	Scalar	Get value of a hash field
redis_hset(key, field, value, secret)	Scalar	Set value of a hash field
redis_lpush(key, value, secret)	Scalar	Push value to a list
redis_lrange(key, start, stop, secret)	Scalar	Get range from a list (comma-separated)
redis_del(key, secret)	Scalar	Delete a key (returns TRUE if deleted)
redis_exists(key, secret)	Scalar	Check if a key exists (returns TRUE if exists)
redis_type(key, secret)	Scalar	Get the type of a key
redis_scan(cursor, pattern, count, secret)	Scalar	Scan keys (returns cursor:keys_csv)
redis_hscan(key, cursor, pattern, count, secret)	Scalar	Scan fields in a hash
redis_keys(pattern, secret)	Table	List all keys matching a pattern
redis_hgetall(key, secret)	Table	List all fields and values in a hash
redis_lrange_table(key, start, stop, secret)	Table	List elements in a list as rows
redis_hscan_over_scan(scan_pattern, hscan_pattern, count, secret)	Table	For all keys matching scan_pattern, HSCAN with hscan_pattern, return (key, field, value) rows

Installation

INSTALL redis FROM community;
LOAD redis;

Usage

Setting up Redis Connection

First, create a secret to store your Redis connection details:

-- Create a Redis connection secret
CREATE SECRET IF NOT EXISTS redis (
        TYPE redis,
        PROVIDER config,
        host 'localhost',
        port '6379',
        password 'optional_password'
    );

-- Create a Redis cloud connection secret
CREATE SECRET IF NOT EXISTS redis (
        TYPE redis,
        PROVIDER config,
        host 'redis-1234.ec2.redns.redis-cloud.com',
        port '16959',
        password 'xxxxxx'
    );

String Operations

-- Set a value
SELECT redis_set('user:1', 'John Doe', 'redis') as result;

-- Get a value
SELECT redis_get('user:1', 'redis') as user_name;

-- Set multiple values in a query
INSERT INTO users (id, name)
SELECT id, redis_set(
    'user:' || id::VARCHAR,
    name,
    'my_redis'
)
FROM new_users;

Hash Operations

-- Set hash fields
SELECT redis_hset('user:1', 'email', 'john@example.com', 'redis');
SELECT redis_hset('user:1', 'age', '30', 'redis');

-- Get hash field
SELECT redis_hget('user:1', 'email', 'redis') as email;

-- Get all fields and values in a hash (table)
SELECT * FROM redis_hgetall('user:1', 'redis');

-- HSCAN a hash (pattern match fields)
SELECT redis_hscan('user:1', '0', 'email*', 100, 'redis');

-- HSCAN over SCAN: for all keys matching a pattern, get all hash fields matching a pattern
SELECT * FROM redis_hscan_over_scan('user:*', 'email*', 100, 'redis');

-- Store user profile as hash
WITH profile(id, field, value) AS (
    VALUES 
        (1, 'name', 'John Doe'),
        (1, 'email', 'john@example.com'),
        (1, 'age', '30')
)
SELECT redis_hset(
    'user:' || id::VARCHAR,
    field,
    value,
    'redis'
)
FROM profile;

List Operations

-- Push items to list
SELECT redis_lpush('mylist', 'first_item', 'redis');
SELECT redis_lpush('mylist', 'second_item', 'redis');

-- Get range from list (returns comma-separated values)
SELECT redis_lrange('mylist', 0, -1, 'redis') as items;

-- Get all items in a list as rows (table)
SELECT * FROM redis_lrange_table('mylist', 0, -1, 'redis');

-- Push multiple items
WITH items(value) AS (
    VALUES ('item1'), ('item2'), ('item3')
)
SELECT redis_lpush('mylist', value, 'redis')
FROM items;

Key Operations

-- List all keys matching a pattern (table)
SELECT * FROM redis_keys('user:*', 'redis');

-- Delete a key
SELECT redis_del('user:1', 'redis');

-- Check if a key exists
SELECT redis_exists('user:1', 'redis');

-- Get the type of a key
SELECT redis_type('user:1', 'redis');

Batch and Discovery Operations

-- Get multiple keys at once
SELECT redis_mget('key1,key2,key3', 'redis') as values;

-- Scan keys matching a pattern
SELECT redis_scan('0', 'user:*', 10, 'redis') as result;

-- Scan all keys matching a pattern (recursive)
WITH RECURSIVE scan(cursor, keys) AS (
    SELECT split_part(redis_scan('0', 'user:*', 10, 'redis'), ':', 1),
           split_part(redis_scan('0', 'user:*', 10, 'redis'), ':', 2)
    UNION ALL
    SELECT split_part(redis_scan(cursor, 'user:*', 10, 'redis'), ':', 1),
           split_part(redis_scan(cursor, 'user:*', 10, 'redis'), ':', 2)
    FROM scan
    WHERE cursor != '0'
)
SELECT keys FROM scan;

Error Handling

The extension functions will throw exceptions with descriptive error messages when:

• Redis secret is not found or invalid
• Unable to connect to Redis server
• Network communication errors occur
• Invalid Redis protocol responses are received

Building from Source

Follow the standard DuckDB extension build process:

# Install vcpkg dependencies
./vcpkg/vcpkg install boost-asio

# Build the extension
make

Future Enhancements

Planned features include:

• Table functions for scanning Redis keys
• Additional Redis commands (SADD, SMEMBERS, etc.)
• Batch operations using Redis pipelines
• Connection timeout handling