Nodes represent entities in your data model (like Person, Product, Company). They can have:

• Labels to categorize them (e.g., :Person, :Movie)
• Properties as key-value pairs
  Relationships connect nodes and represent how entities are related. They:
• Always have a direction (though can be traversed both ways)
• Must have exactly one type (e.g., :ACTED_IN, :FRIENDS_WITH)
• Can have properties
  Properties are key-value pairs that store data on both nodes and relationships.
  Example:

(person:Person {name: "Tom Hanks", born: 1956})
-[:ACTED_IN {role: "Forrest"}]->
(movie:Movie {title: "Forrest Gump", released: 1994})

Cypher is Neo4j's declarative query language designed specifically for working with graph data. Key differences from SQL:
Cypher characteristics:

• Uses ASCII art to represent graph patterns: (a)-[:REL]->(b)
• Focuses on pattern matching rather than table joins
• More intuitive for expressing graph traversals
• Supports path expressions naturally
  Example comparison:

// Cypher: Find friends of friends
MATCH (me:Person {name: "Alice"})-[:FRIEND]-(friend)-[:FRIEND]-(fof)
RETURN fof.name
// SQL equivalent would require multiple JOINs and be more complex

Creating nodes:

CREATE (p:Person {name: "John", age: 30})

Creating relationships:

// Between existing nodes
MATCH (a:Person {name: "John"}), (b:Person {name: "Jane"})
CREATE (a)-[:KNOWS {since: 2020}]->(b)
// Create nodes and relationship together
CREATE (a:Person {name: "Bob"})-[:WORKS_FOR]->(c:Company {name: "TechCorp"})

Best practices:

• Use MERGE instead of CREATE to avoid duplicates
• Always specify node labels for better performance
• Use meaningful relationship types

CREATE always creates new nodes/relationships, potentially causing duplicates:

CREATE (p:Person {name: "John"})  // Always creates new node

MERGE finds existing patterns or creates them if they don't exist:

MERGE (p:Person {name: "John"})   // Creates only if doesn't exist

MERGE with ON CREATE/MATCH:

MERGE (p:Person {email: "john@example.com"})
ON CREATE SET p.created = timestamp(), p.name = "John"
ON MATCH SET p.lastSeen = timestamp()

MERGE is safer for avoiding duplicates but can be slower than CREATE when you're certain the data doesn't exist.