Hash vs string for objects in Redis - Performance Comparison

Time Complexity: Hash vs string for objects

O(n) for string whole-object access, O(1) for hash single field access

Understanding Time Complexity

When storing objects in Redis, we can use either strings or hashes. Understanding how the time to access or update data changes with size helps us choose the best option.

We want to know how the time to get or set fields grows as the object gets bigger.

Scenario Under Consideration

Analyze the time complexity of these Redis commands for storing and accessing objects.

# Using string to store whole object
SET user:1000 "{\"name\":\"Alice\", \"age\":30, \"city\":\"NY\"}"
GET user:1000

# Using hash to store fields separately
HSET user:1000 name "Alice" age 30 city "NY"
HGET user:1000 name
HGETALL user:1000
    

This code shows storing an object as one string versus storing fields in a hash and accessing them.

Identify Repeating Operations

Look at what repeats when accessing or updating data.

• Primary operation: Accessing fields inside the stored object.
• How many times: For strings, the whole object is read or written at once. For hashes, each field is accessed individually.

How Execution Grows With Input

Think about how time changes as the object grows with more fields.

Input Size (fields)	Approx. Operations
10	String: reads/writes whole object (small size), Hash: reads/writes individual fields (10 ops for all fields)
100	String: reads/writes whole bigger object, Hash: 100 ops for all fields, 1 op for single field
1000	String: reads/writes very large object, Hash: 1000 ops for all fields, 1 op for single field

Pattern observation: String operations take longer as the whole object grows. Hash single field access stays fast regardless of object size.

Final Time Complexity

Time Complexity: O(n) for string whole-object access, O(1) for hash single field access

This means reading or writing the whole string grows with object size, but accessing one field in a hash stays quick no matter how big the object is.

Common Mistake

[X] Wrong: "Accessing a field in a hash takes as long as reading the whole object stored as a string."

[OK] Correct: Hashes let Redis jump directly to the field, so single field access is fast and does not depend on total object size.

Interview Connect

Knowing how Redis handles strings and hashes helps you pick the right data structure for fast access. This skill shows you understand practical trade-offs in real projects.

Self-Check

"What if we used HGETALL to get all fields from a hash instead of HGET for one field? How would the time complexity change?"

Practice

1. In Redis, when should you prefer using a hash over a string to store data?

easy

A. When you want to store multiple related fields under one key, like an object.

B. When you only need to store a single simple value per key.

C. When you want to store large binary data efficiently.

D. When you want to store data that changes very rarely.

Solution

1. Step 1: Understand Redis data types

   Strings store single values per key, while hashes can store multiple fields under one key.

2. Step 2: Match use case to data type

   If you want to store an object with multiple fields (like name, age, email), hashes are better suited.

3. Final Answer:

   When you want to store multiple related fields under one key, like an object. -> Option A

4. Quick Check:

   Hash = multiple fields per key [OK]

Hint: Use hashes for objects with many fields, strings for single values [OK]

Common Mistakes:

• Thinking strings can store multiple fields easily
• Using hashes for simple single values
• Confusing hashes with lists or sets

2. Which of the following Redis commands correctly sets a field age to 30 in a hash named user:1?

easy

A. SET user:1 age 30

B. HSET user:1 age 30

C. HMSET user:1 age: 30

D. SETFIELD user:1 age 30

Solution

1. Step 1: Identify command for setting a single field in a hash

   The HSET command sets a field in a hash key.

2. Step 2: Check other options

   SET is for strings, HMSET is deprecated but similar to HSET, SETFIELD is invalid.

3. Final Answer:

   HSET user:1 age 30 -> Option B

4. Quick Check:

   HSET sets hash fields [OK]

Hint: Use HSET to set one field in a hash [OK]

Common Mistakes:

• Using SET for hash fields
• Using deprecated HMSET instead of HSET
• Inventing commands like SETFIELD

3. Given the following Redis commands executed in order:

HSET user:2 name "Alice" age 25
HSET user:2 city "Paris"
GET user:2

What will be the output of the GET user:2 command?

medium

A. null (no value)

B. "Alice"

C. A list of all fields and values in user:2.

D. An error because user:2 is a hash, not a string.

Solution

1. Step 1: Understand data types and commands

   HSET creates a hash key user:2. GET works only on string keys.

2. Step 2: Predict command behavior

   Using GET on a hash key causes a type error in Redis.

3. Final Answer:

   An error because user:2 is a hash, not a string. -> Option D

4. Quick Check:

   GET on hash key = error [OK]

Hint: GET only works on strings, not hashes [OK]

Common Mistakes:

• Expecting GET to return hash fields
• Confusing GET with HGETALL
• Assuming null means no error

4. You have this Redis command sequence:

HSET user:3 name "Bob" age 40
SET user:3 "simple string"

What problem will occur and how to fix it?

medium

A. Error because user:3 key type conflicts; delete key before SET.

B. No problem; Redis allows both hash and string on same key.

C. The SET command will overwrite the hash silently.

D. The HSET command will fail because key exists as string.

Solution

1. Step 1: Understand Redis key type rules

   A key can only hold one data type at a time. Setting a string on a key that is a hash causes an error.

2. Step 2: Identify fix

   To set a string on the same key, delete the existing hash key first using DEL.

3. Final Answer:

   Error because user:3 key type conflicts; delete key before SET. -> Option A

4. Quick Check:

   Key type conflict = error, DEL key first [OK]

Hint: Delete key before changing type to avoid errors [OK]

Common Mistakes:

• Assuming Redis allows multiple types on one key
• Expecting SET to overwrite hash without error
• Not deleting key before changing type

5. You want to store user profiles with fields name, age, and email in Redis. You also want to efficiently update individual fields without rewriting the whole object. Which approach is best and why?

hard

A. Store each user as multiple string keys, one per field, like user:1:name, user:1:age.

B. Store each user as a string with JSON and rewrite entire string on update.

C. Store each user as a hash with fields for name, age, and email for easy updates.

D. Store all users in one big hash with user IDs as fields and JSON strings as values.

Solution

1. Step 1: Consider update efficiency

   Updating individual fields is easiest with hashes because you can update one field without rewriting others.

2. Step 2: Compare options

   Strings with JSON require rewriting whole string. Multiple string keys increase complexity. One big hash mixes users and complicates access.

3. Final Answer:

   Store each user as a hash with fields for name, age, and email for easy updates. -> Option C

4. Quick Check:

   Hashes = efficient field updates [OK]

Hint: Use hashes for objects needing partial field updates [OK]

Common Mistakes:

• Using strings with JSON and rewriting whole data
• Splitting fields into many string keys unnecessarily
• Storing all users in one hash causing complexity