Working with Key–Value Databases

Key–value databases store data as pairs of a unique key and a value.
This is similar to a Python dictionary or a JavaScript object.

In this workshop, we will use Redis, a popular open-source key–value database.

Setup Redis

Setting up Redis locally

You can run Redis locally using Docker:

docker run -d --name redis-demo -p 6379:6379 redis

Connect to the Redis command-line interface:

docker exec -it redis-demo redis-cli

You should see a prompt like this:

127.0.0.1:6379>

Execute Redis commands in a browser

Via the OneCompiler website it is possible to execute Redis commands in a virtual enviroment. The website address is: https://onecompiler.com/redis

💡 Quoting values with spaces

When using Redis in an online environment such as OneCompiler, always use single quotes (' ') around values that contain spaces.

Example:
HSET product:1002 name 'MacBook Air' price 1299 brand 'Apple'
If you use double quotes (" ") instead, the platform may split "MacBook Air" into two separate arguments (MacBook and Air) and Redis will return:
(error) WRONG NUMBER OF ARGUMENTS FOR 'hset' command
Using single quotes ensures the entire string is treated as one value.

Basic Commands

Storing and retrieving values

SET product:1001 "Laptop"
GET product:1001

Redis stores the value “Laptop” under the key “product:1001”.

Updating values

SET product:1001 "Laptop – Dell XPS 13"
GET product:1001

The SET command overwrites the existing value for the same key.

About key naming conventions

In Redis, a key such as product:1005 is not a real namespace — it is simply a naming convention.

Redis stores all keys in a single global keyspace.
The colon (:) has no technical meaning; it is just part of the key name.

Developers use the : convention to make keys easier to read and group logically:

Example Key Meaning

product:1005 Product with ID 1005

cart:alice Shopping cart belonging to user Alice

session:uid123 Session for user with ID 123

This makes it easy to filter related keys:
KEYS product:*

Example Key	Meaning
`product:1005`	Product with ID 1005
`cart:alice`	Shopping cart belonging to user Alice
`session:uid123`	Session for user with ID 123

Using hashes for structured data
Instead of a single string, you can store multiple fields using a hash:

HSET product:1002 name "MacBook Air" price 1299 brand "Apple"

The value of the key product:1002 is a mapping on its own:

 product:1002 = {
  "name": "MacBook Air",
  "price": "1299",
  "brand": "Apple"
}

To get all values from this mapping:

HGETALL product:1002

Output:

1) "name"
2) "MacBook Air"
3) "price"
4) "1299"
5) "brand"
6) "Apple"

To get a specific field from this mapping:

HGET product:1002 price

Output:

Lists
Lists store ordered values. You can use them as queues or logs.

RPUSH cart:alice "book_123"
RPUSH cart:alice "laptop_225"
RPUSH cart:alice "mouse_25"
LRANGE cart:alice 0 -1

The RPUSH command stands for Right push, so it pushes a element at the end of the list. The LPUSH command pushes a element at the start of the list.
The LRANGE command returns elements within a given range. It takes two arguments: the start and stop index. 0 represents the first element, -1 the last element.

The result:

1) "book_123"
2) "laptop_225"
3) "mouse_25"

graph LR
    subgraph "Redis List: cart:alice"
        A["index: 0<br/>book_123"] --> B["index: 1<br/>laptop_225"] --> C["index: 2<br/>mouse_25"]
    end
   
    E["LRANGE 0 -1"] --> A

Expiring Keys
You can make a key temporary — useful for sessions or caches.

SET session:alice "logged_in"
EXPIRE session:alice 60
TTL session:alice

This sets a time-to-live of 60 seconds for the key.

Practice Questions

🧠 1. Create a new hash for a product “product:2001” with the following fields:

name = "Smartphone", brand = "Samsung", price = 799.

Click to reveal answer

HSET product:2001 name "Smartphone" brand "Samsung" price 799 HGETALL product:2001

🧠 2. Use a list to simulate a shopping cart for Bob with three items:

book_101, laptop_225, and phone_331.

Retrieve all items from his cart.

Click to reveal answer

LPUSH cart:bob "book_101" "laptop_225" "phone_331" LRANGE cart:bob 0 -1

🧠 3. Set a temporary key for “discount:today” with the value “15%”, expiring after 120 seconds.

Click to reveal answer

SET discount:today "15%" EXPIRE discount:today 120 TTL discount:today

🧠

How could you store user login sessions with automatic expiry?

Click to reveal discussion

You can use one key per session ID, e.g.: \ `SET session:uid123 "active" EXPIRE session:uid123 3600` \ This key expires automatically after one hour.

Searching in Key–Value Databases

Key–value databases such as Redis are optimized for speed, not for complex queries.
They can retrieve a value extremely fast — but only if you already know the key.

For example:

GET product:1005

returns the value immediately, because Redis looks up the key directly in memory. This makes access times almost instantaneous. However, you cannot search inside the value. Redis has no concept of “fields” or “columns” like an SQL or document database.

If you store:

HSET product:1005 name "Laptop" brand "Dell" price 899

you cannnot ask:

# Not possible
FIND all products WHERE brand = "Dell"

Redis doesn’t understand what a brand or price field means — it simply stores bytes under the key product:1005.

If you need to search by field values (e.g. all products by brand), you must model such queries yourself, for example by creating additional index-like structures:

SADD brand:Dell product:1005
SADD brand:Apple product:1006
SMEMBERS brand:Dell

This returns all keys for Dell products. The commands SADD and SMEMBERS are used to work with sets. A set is a collection of unique unordered elements. We will discuss sets further in the next chapter.