Redis hashes are maps between string fields and string values. They’re perfect for representing objects and provide efficient field-level operations without fetching the entire object.
Use Cases
- User profiles: Store user attributes (name, email, age)
- Product catalogs: Product properties (name, price, stock)
- Session storage: Multiple session attributes
- Object caching: Serialize objects as hashes
- Counters: Multiple related counters
- Form data: Temporary storage of form submissions
Key Commands
Basic Operations
# Set fields
redis> HSET user:123 name "Alice" email "[email protected]" age "30"
(integer) 3
# Get single field
redis> HGET user:123 name
"Alice"
# Get all fields and values
redis> HGETALL user:123
1) "name"
2) "Alice"
3) "email"
4) "[email protected]"
5) "age"
6) "30"
# Get multiple fields
redis> HMGET user:123 name email
1) "Alice"
2) "[email protected]"
# Check if field exists
redis> HEXISTS user:123 name
(integer) 1
Field Operations
# Set only if field doesn't exist
redis> HSETNX user:123 status "active"
(integer) 1
redis> HSETNX user:123 status "inactive"
(integer) 0
# Delete fields
redis> HDEL user:123 age
(integer) 1
# Get all field names
redis> HKEYS user:123
1) "name"
2) "email"
3) "status"
# Get all values
redis> HVALS user:123
1) "Alice"
2) "[email protected]"
3) "active"
# Get field count
redis> HLEN user:123
(integer) 3
Numeric Operations
# Increment by integer
redis> HINCRBY stats:page:123 views 1
(integer) 1
redis> HINCRBY stats:page:123 views 5
(integer) 6
# Increment by float
redis> HINCRBYFLOAT product:456 price 0.50
"19.99"
Hash Field Expiration (HFE)
Redis supports TTL on individual hash fields:
# Set field with expiration (seconds)
redis> HSET session:xyz token "abc123"
(integer) 1
redis> HEXPIRE session:xyz 3600 FIELDS 1 token
1) (integer) 1
# Check field TTL
redis> HTTL session:xyz FIELDS 1 token
1) (integer) 3599
# Persist field (remove expiration)
redis> HPERSIST session:xyz FIELDS 1 token
1) (integer) 1
Hash Field Expiration is available in Redis 7.4+ and allows setting TTL on individual fields within a hash.
Time Complexity
| Command | Time Complexity | Description |
|---|
| HSET | O(1) per field | Set field value |
| HGET | O(1) | Get field value |
| HMGET | O(N) | N=fields requested |
| HDEL | O(N) | N=fields deleted |
| HGETALL | O(N) | N=hash size |
| HKEYS/HVALS | O(N) | N=hash size |
| HEXISTS | O(1) | Check field existence |
| HLEN | O(1) | Get field count |
| HINCRBY | O(1) | Increment field |
Patterns and Examples
User Profile
# Store user data
redis> HSET user:1001 username "alice" email "[email protected]" \
created "2024-03-03" reputation "1523"
(integer) 4
# Update single field
redis> HSET user:1001 reputation "1550"
(integer) 0
# Increment reputation
redis> HINCRBY user:1001 reputation 10
(integer) 1560
# Get user data
redis> HGETALL user:1001
1) "username"
2) "alice"
3) "email"
4) "[email protected]"
5) "created"
6) "2024-03-03"
7) "reputation"
8) "1560"
Product Catalog
# Store product
redis> HSET product:2001 name "Redis Book" price "29.99" \
stock "150" category "books"
(integer) 4
# Update stock after purchase
redis> HINCRBY product:2001 stock -1
(integer) 149
# Check stock before purchase
redis> HGET product:2001 stock
"149"
# Update price
redis> HSET product:2001 price "24.99"
(integer) 0
Session Storage
# Create session
redis> HSET session:abc123 user_id "1001" ip "192.168.1.1" \
created "1709467200" last_activity "1709467200"
(integer) 4
# Update last activity
redis> HSET session:abc123 last_activity "1709467260"
(integer) 0
# Set session expiration
redis> EXPIRE session:abc123 3600
(integer) 1
# Get session data
redis> HMGET session:abc123 user_id ip
1) "1001"
2) "192.168.1.1"
Statistics Tracking
# Track page statistics
redis> HINCRBY stats:page:home views 1
(integer) 1523
redis> HINCRBY stats:page:home unique_visitors 1
(integer) 891
# Get all stats
redis> HGETALL stats:page:home
1) "views"
2) "1523"
3) "unique_visitors"
4) "891"
# Calculate conversion rate
redis> HMGET stats:page:home views conversions
1) "1523"
2) "45"
# Add items to cart (item_id: quantity)
redis> HSET cart:user:123 item:456 "2" item:789 "1"
(integer) 2
# Update quantity
redis> HINCRBY cart:user:123 item:456 1
(integer) 3
# Remove item
redis> HDEL cart:user:123 item:789
(integer) 1
# Get cart contents
redis> HGETALL cart:user:123
1) "item:456"
2) "3"
# Set cart expiration (e.g., 7 days)
redis> EXPIRE cart:user:123 604800
(integer) 1
Internal Encoding
Hashes use two encodings:
- listpack: Small hashes (memory efficient)
- hashtable: Larger hashes (faster field access)
redis> HSET small a "1" b "2" c "3"
(integer) 3
redis> OBJECT ENCODING small
"listpack"
redis> HSET large field1 "value1"
# ... add many fields
redis> OBJECT ENCODING large
"hashtable"
# Max listpack entries/value before converting to hashtable
hash-max-listpack-entries 512
hash-max-listpack-value 64
Best Practices
- Use hashes for objects - more memory efficient than separate keys
- Store related counters in a single hash
- Set expiration on the hash key for automatic cleanup
- Use HMGET instead of multiple HGET calls
- Avoid HGETALL on large hashes - use HSCAN for iteration
HGETALL returns all fields and values. For large hashes (over 100 fields), use HSCAN to iterate incrementally.
Hashes are more memory efficient than using separate keys. For example, user:123:name, user:123:email use more memory than HSET user:123 name email.
Hashes vs Other Types
| Need | Use | Why |
|---|
| Single value | String | Simplest |
| Multiple related values | Hash | Efficient, field-level ops |
| Complex nested objects | JSON | Native JSON querying |
| List of values | List | Ordered collection |
| Unique values | Set | Uniqueness guarantee |
Memory Comparison
# Separate keys: ~100 bytes overhead per key
redis> SET user:123:name "Alice"
redis> SET user:123:email "[email protected]"
redis> SET user:123:age "30"
# Total: ~300 bytes overhead
# Single hash: ~100 bytes overhead total
redis> HSET user:123 name "Alice" email "[email protected]" age "30"
# Total: ~100 bytes overhead (3x more efficient)
Next Steps
Hash Commands
Complete command reference
JSON
For complex nested objects