Amazon DynamoDB is a fully managed NoSQL database service provided by AWS. Key characteristics include:

• Serverless: No infrastructure management required
• High Performance: Single-digit millisecond latency at any scale
• Highly Available: 99.99% availability SLA with multi-AZ replication
• Flexible Schema: NoSQL document and key-value data models
• Auto Scaling: Automatically scales up/down based on traffic
• ACID Transactions: Support for transactions across multiple items
• Global Tables: Multi-region, multi-master replication

DynamoDB supports several data types:

Scalar Types:

• String (S)
• Number (N)
• Binary (B)
• Boolean (BOOL)
• Null (NULL)

Collection Types:

• String Set (SS)
• Number Set (NS)
• Binary Set (BS)
• List (L) - ordered collection of values
• Map (M) - unordered collection of key-value pairs

A partition key is the primary key component that determines which physical partition your data is stored on. It's crucial because:

• Data Distribution: DynamoDB uses the partition key value to distribute data across multiple partitions
• Performance: Even distribution prevents hot partitions
• Scalability: Enables horizontal scaling across partitions

Example: In a user table, user_id as partition key distributes users evenly across partitions.

Query:

• Uses primary key or index keys
• Retrieves specific items efficiently
• Lower cost and better performance
• Can use filter expressions for additional filtering

Scan:

• Examines every item in table/index
• Expensive and slow for large tables
• Uses filter expressions to return subset
• Should be avoided in production for large datasets

# Query example
response = table.query(
    KeyConditionExpression=Key('user_id').eq('123')
)

# Scan example  
response = table.scan(
    FilterExpression=Attr('age').gt(21)
)