# Neo4j Python Driver

The Neo4j Python Driver is the official driver for connecting Python applications to Neo4j graph databases using the Bolt protocol. It provides both synchronous and asynchronous APIs for executing Cypher queries, managing transactions, and handling results. The driver supports Python 3.10+ and works with Neo4j server versions through automatic protocol negotiation.

The driver offers multiple transaction patterns including auto-commit transactions, explicit transactions, and managed transaction functions with automatic retry logic for transient errors. It supports connection pooling, causal consistency through bookmarks, cluster routing for high availability deployments, and provides type mappings for Neo4j's spatial, temporal, and graph data types.

## Creating a Driver Connection

Create a driver instance to establish a connection to a Neo4j database. The driver manages connection pooling and should be reused throughout the application lifecycle.

```python
from neo4j import GraphDatabase

# Basic connection with authentication
URI = "neo4j://localhost:7687"
AUTH = ("neo4j", "password")

driver = GraphDatabase.driver(URI, auth=AUTH)

# Verify connectivity
driver.verify_connectivity()

# Get server information
server_info = driver.get_server_info()
print(f"Connected to: {server_info.agent}")
print(f"Protocol version: {server_info.protocol_version}")

# Always close the driver when done
driver.close()

# Or use as context manager (recommended)
with GraphDatabase.driver(URI, auth=AUTH) as driver:
    driver.verify_connectivity()
    # Use driver here...
# Driver automatically closed
```

## Authentication Methods

Configure authentication using built-in helper functions for various authentication schemes.

```python
from neo4j import GraphDatabase, basic_auth, bearer_auth, kerberos_auth, custom_auth

# Basic authentication (username/password)
auth = basic_auth("neo4j", "password")
driver = GraphDatabase.driver("neo4j://localhost:7687", auth=auth)

# Tuple shorthand for basic auth
driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

# Bearer token authentication (SSO)
auth = bearer_auth("base64_encoded_token")
driver = GraphDatabase.driver("neo4j://localhost:7687", auth=auth)

# Kerberos authentication
auth = kerberos_auth("base64_encoded_ticket")
driver = GraphDatabase.driver("neo4j://localhost:7687", auth=auth)

# Custom authentication scheme
auth = custom_auth(
    principal="neo4j",
    credentials="password",
    realm="native",
    scheme="basic",
    custom_param="value"
)
driver = GraphDatabase.driver("neo4j://localhost:7687", auth=auth)

driver.close()
```

## Driver Configuration Options

Configure the driver with various options for connection management, timeouts, and encryption.

```python
from neo4j import GraphDatabase, TrustSystemCAs, TrustAll, TrustCustomCAs

# Full driver configuration
driver = GraphDatabase.driver(
    "neo4j://localhost:7687",
    auth=("neo4j", "password"),
    # Connection pool settings
    max_connection_pool_size=100,
    max_connection_lifetime=3600,  # seconds
    connection_acquisition_timeout=60,  # seconds
    # Connection settings
    connection_timeout=30,  # seconds
    # Retry settings
    max_transaction_retry_time=30,  # seconds
    # Encryption (for bolt:// scheme)
    encrypted=True,
    trusted_certificates=TrustSystemCAs(),  # or TrustAll(), TrustCustomCAs(["cert.pem"])
    # Custom user agent
    user_agent="MyApp/1.0",
)

# Using secure URI schemes (automatic encryption)
# bolt+s:// - encrypted with system CAs
# bolt+ssc:// - encrypted with self-signed certificates
# neo4j+s:// - routing + encrypted with system CAs
# neo4j+ssc:// - routing + encrypted with self-signed certificates

driver_secure = GraphDatabase.driver(
    "neo4j+s://localhost:7687",
    auth=("neo4j", "password")
)

driver.close()
driver_secure.close()
```

## execute_query - Simple Query Execution

The simplest way to run queries with automatic session and transaction management.

```python
from neo4j import GraphDatabase, RoutingControl

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

# Simple write query
records, summary, keys = driver.execute_query(
    "CREATE (p:Person {name: $name, age: $age}) RETURN p",
    name="Alice",
    age=30,
    database_="neo4j"
)
print(f"Created {summary.counters.nodes_created} nodes")

# Read query with routing control
records, summary, keys = driver.execute_query(
    "MATCH (p:Person) WHERE p.age > $min_age RETURN p.name AS name, p.age AS age",
    min_age=25,
    database_="neo4j",
    routing_=RoutingControl.READ  # Route to read replica
)
for record in records:
    print(f"{record['name']} is {record['age']} years old")

# Using result transformer for custom return types
from neo4j import Result

record = driver.execute_query(
    "MATCH (p:Person {name: $name}) RETURN p.age AS age",
    name="Alice",
    result_transformer_=Result.single,  # Returns single Record or None
    database_="neo4j"
)
if record:
    print(f"Age: {record['age']}")

driver.close()
```

## Session Management

Sessions provide the context for executing transactions and managing database state.

```python
from neo4j import GraphDatabase, READ_ACCESS, WRITE_ACCESS

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

# Basic session usage with context manager
with driver.session(database="neo4j") as session:
    result = session.run("MATCH (n) RETURN count(n) AS count")
    count = result.single()["count"]
    print(f"Node count: {count}")

# Session with configuration
with driver.session(
    database="neo4j",
    default_access_mode=READ_ACCESS,
    fetch_size=1000,  # Records fetched per batch
    impersonated_user="other_user"  # Run as different user
) as session:
    result = session.run("MATCH (n:Person) RETURN n.name AS name")
    names = [record["name"] for record in result]

# Getting bookmarks for causal consistency
with driver.session(database="neo4j") as session:
    session.run("CREATE (p:Person {name: 'Bob'})").consume()
    bookmarks = session.last_bookmarks()

# Using bookmarks in subsequent session
with driver.session(database="neo4j", bookmarks=bookmarks) as session:
    # This session will see the changes from the previous session
    result = session.run("MATCH (p:Person {name: 'Bob'}) RETURN p")
    record = result.single()

driver.close()
```

## Auto-commit Transactions (session.run)

Execute single queries directly through the session. Best for simple operations that don't require transaction control.

```python
from neo4j import GraphDatabase, Query

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

with driver.session(database="neo4j") as session:
    # Simple query execution
    result = session.run(
        "CREATE (p:Person {name: $name}) RETURN p.name AS name",
        name="Charlie"
    )
    record = result.single()
    print(f"Created: {record['name']}")

    # Query with timeout using Query object
    result = session.run(
        Query("MATCH (n) RETURN count(n) AS count", timeout=5.0)
    )
    print(f"Count: {result.single()['count']}")

    # Parameterized query with kwargs
    result = session.run(
        "MATCH (p:Person) WHERE p.name = $name AND p.age > $age RETURN p",
        name="Alice",
        age=25
    )
    for record in result:
        print(record["p"])

    # Consume result to get summary
    result = session.run("CREATE (n:Test)")
    summary = result.consume()
    print(f"Query completed in {summary.result_available_after} ms")
    print(f"Nodes created: {summary.counters.nodes_created}")

driver.close()
```

## Managed Transaction Functions (execute_read/execute_write)

Use transaction functions for automatic retry on transient errors. The transaction is committed if the function succeeds, rolled back if it raises an exception.

```python
from neo4j import GraphDatabase, ManagedTransaction, unit_of_work

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

# Define transaction functions
def create_person(tx: ManagedTransaction, name: str, age: int) -> str:
    result = tx.run(
        "CREATE (p:Person {name: $name, age: $age}) RETURN p.name AS name",
        name=name,
        age=age
    )
    return result.single()["name"]

def get_people_by_age(tx: ManagedTransaction, min_age: int) -> list[dict]:
    result = tx.run(
        "MATCH (p:Person) WHERE p.age >= $min_age RETURN p.name AS name, p.age AS age",
        min_age=min_age
    )
    return [{"name": record["name"], "age": record["age"]} for record in result]

# Execute write transaction
with driver.session(database="neo4j") as session:
    name = session.execute_write(create_person, "David", 35)
    print(f"Created: {name}")

# Execute read transaction
with driver.session(database="neo4j") as session:
    people = session.execute_read(get_people_by_age, 30)
    for person in people:
        print(f"{person['name']}: {person['age']}")

# Transaction function with timeout and metadata using decorator
@unit_of_work(timeout=10.0, metadata={"app": "my_app"})
def create_with_timeout(tx: ManagedTransaction, name: str):
    result = tx.run("CREATE (p:Person {name: $name}) RETURN p", name=name)
    return result.single()["p"]

with driver.session(database="neo4j") as session:
    person = session.execute_write(create_with_timeout, "Eve")

driver.close()
```

## Explicit Transactions (begin_transaction)

Full control over transaction lifecycle for complex operations requiring manual commit/rollback.

```python
from neo4j import GraphDatabase

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

# Using transaction as context manager (auto-commit on success, rollback on exception)
with driver.session(database="neo4j") as session:
    with session.begin_transaction() as tx:
        tx.run("CREATE (a:Person {name: 'Frank'})")
        tx.run("CREATE (b:Person {name: 'Grace'})")
        tx.run(
            "MATCH (a:Person {name: 'Frank'}), (b:Person {name: 'Grace'}) "
            "CREATE (a)-[:KNOWS]->(b)"
        )
        # Transaction commits automatically when exiting with block

# Manual transaction control
with driver.session(database="neo4j") as session:
    tx = session.begin_transaction(
        metadata={"operation": "batch_import"},
        timeout=30.0
    )
    try:
        tx.run("CREATE (n:Node {id: 1})")
        tx.run("CREATE (n:Node {id: 2})")
        # Check some condition
        result = tx.run("MATCH (n:Node) RETURN count(n) AS count")
        count = result.single()["count"]
        if count >= 2:
            tx.commit()
            print("Transaction committed")
        else:
            tx.rollback()
            print("Transaction rolled back")
    except Exception as e:
        tx.rollback()
        raise

driver.close()
```

## Working with Results

Process query results with various methods for different use cases.

```python
from neo4j import GraphDatabase

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

with driver.session(database="neo4j") as session:
    # Setup test data
    session.run("MATCH (n) DETACH DELETE n").consume()
    session.run(
        "UNWIND range(1, 5) AS i CREATE (p:Person {name: 'Person' + toString(i), age: 20 + i})"
    ).consume()

    # Iterate over results
    result = session.run("MATCH (p:Person) RETURN p.name AS name, p.age AS age")
    for record in result:
        print(f"{record['name']}: {record['age']}")

    # Get single record (warns if multiple, None if empty)
    result = session.run("MATCH (p:Person {name: 'Person1'}) RETURN p.name AS name")
    record = result.single()
    if record:
        print(f"Found: {record['name']}")

    # Get single record (strict - raises if not exactly one)
    result = session.run("MATCH (p:Person {name: 'Person1'}) RETURN p.name AS name")
    record = result.single(strict=True)
    print(f"Found: {record['name']}")

    # Fetch specific number of records
    result = session.run("MATCH (p:Person) RETURN p.name AS name ORDER BY p.name")
    first_two = result.fetch(2)
    print(f"First two: {[r['name'] for r in first_two]}")

    # Peek at next record without consuming
    result = session.run("MATCH (p:Person) RETURN p.name AS name ORDER BY p.name")
    next_record = result.peek()
    print(f"Next will be: {next_record['name']}")

    # Get values as list
    result = session.run("MATCH (p:Person) RETURN p.name AS name ORDER BY p.name")
    names = result.value("name")  # List of name values
    print(f"Names: {names}")

    # Get all values as list of lists
    result = session.run("MATCH (p:Person) RETURN p.name, p.age ORDER BY p.name")
    all_values = result.values()  # [[name1, age1], [name2, age2], ...]

    # Convert to list of dictionaries
    result = session.run("MATCH (p:Person) RETURN p.name AS name, p.age AS age")
    data = result.data()
    print(f"Data: {data}")

    # Get result keys
    result = session.run("MATCH (p:Person) RETURN p.name AS name, p.age AS age")
    print(f"Keys: {result.keys()}")

    # Convert to EagerResult (buffers all records)
    result = session.run("MATCH (p:Person) RETURN p.name AS name")
    eager = result.to_eager_result()
    print(f"Records: {eager.records}, Keys: {eager.keys}")

driver.close()
```

## Result Summary and Counters

Access query execution statistics and metadata through the result summary.

```python
from neo4j import GraphDatabase

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

with driver.session(database="neo4j") as session:
    result = session.run(
        "CREATE (a:Person {name: 'Test1'})-[r:KNOWS]->(b:Person {name: 'Test2'}) "
        "SET a.updated = true "
        "RETURN a, r, b"
    )
    summary = result.consume()

    # Query timing
    print(f"Query available after: {summary.result_available_after} ms")
    print(f"Query consumed after: {summary.result_consumed_after} ms")

    # Counters for write operations
    counters = summary.counters
    print(f"Nodes created: {counters.nodes_created}")
    print(f"Nodes deleted: {counters.nodes_deleted}")
    print(f"Relationships created: {counters.relationships_created}")
    print(f"Relationships deleted: {counters.relationships_deleted}")
    print(f"Properties set: {counters.properties_set}")
    print(f"Labels added: {counters.labels_added}")
    print(f"Labels removed: {counters.labels_removed}")
    print(f"Indexes added: {counters.indexes_added}")
    print(f"Indexes removed: {counters.indexes_removed}")
    print(f"Constraints added: {counters.constraints_added}")
    print(f"Constraints removed: {counters.constraints_removed}")
    print(f"Contains updates: {counters.contains_updates}")
    print(f"Contains system updates: {counters.contains_system_updates}")

    # Server info
    print(f"Server: {summary.server.agent}")
    print(f"Database: {summary.database}")
    print(f"Query type: {summary.query_type}")

    # Query profile (if PROFILE was used)
    result = session.run("PROFILE MATCH (n) RETURN count(n)")
    summary = result.consume()
    if summary.profile:
        print(f"Profile: {summary.profile}")
    if summary.plan:
        print(f"Plan: {summary.plan}")

driver.close()
```

## Working with Graph Types (Node, Relationship, Path)

Handle Neo4j graph entities returned from queries.

```python
from neo4j import GraphDatabase
from neo4j.graph import Node, Relationship, Path

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

with driver.session(database="neo4j") as session:
    # Setup
    session.run("MATCH (n) DETACH DELETE n").consume()
    session.run(
        "CREATE (a:Person:Employee {name: 'Alice', age: 30})"
        "-[:KNOWS {since: 2020}]->"
        "(b:Person {name: 'Bob', age: 25})"
        "-[:KNOWS {since: 2021}]->"
        "(c:Person {name: 'Charlie', age: 35})"
    ).consume()

    # Working with Nodes
    result = session.run("MATCH (p:Person {name: 'Alice'}) RETURN p")
    record = result.single()
    node: Node = record["p"]
    print(f"Element ID: {node.element_id}")
    print(f"Labels: {node.labels}")  # frozenset({'Person', 'Employee'})
    print(f"Properties: {dict(node)}")
    print(f"Name: {node['name']}")
    print(f"Age: {node.get('age', 0)}")

    # Working with Relationships
    result = session.run(
        "MATCH (a:Person {name: 'Alice'})-[r:KNOWS]->(b:Person) RETURN r"
    )
    record = result.single()
    rel: Relationship = record["r"]
    print(f"Relationship type: {rel.type}")
    print(f"Start node: {rel.start_node}")
    print(f"End node: {rel.end_node}")
    print(f"Properties: {dict(rel)}")
    print(f"Since: {rel['since']}")

    # Working with Paths
    result = session.run(
        "MATCH path = (a:Person {name: 'Alice'})-[:KNOWS*]->(c:Person {name: 'Charlie'}) "
        "RETURN path"
    )
    record = result.single()
    path: Path = record["path"]
    print(f"Path length: {len(path)}")
    print(f"Start node: {path.start_node['name']}")
    print(f"End node: {path.end_node['name']}")
    print(f"Nodes in path: {[n['name'] for n in path.nodes]}")
    print(f"Relationships: {[r.type for r in path.relationships]}")

    # Access the graph from results
    result = session.run(
        "MATCH (a:Person)-[r:KNOWS]->(b:Person) RETURN a, r, b"
    )
    graph = result.graph()
    print(f"Nodes in graph: {len(graph.nodes)}")
    print(f"Relationships in graph: {len(graph.relationships)}")
    for node in graph.nodes:
        print(f"  Node: {node['name']}")

driver.close()
```

## Working with Records

Access record data using various methods.

```python
from neo4j import GraphDatabase

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

with driver.session(database="neo4j") as session:
    result = session.run(
        "RETURN 'Alice' AS name, 30 AS age, ['Python', 'Neo4j'] AS skills"
    )
    record = result.single()

    # Access by key name
    print(f"Name: {record['name']}")
    print(f"Age: {record['age']}")

    # Access by index
    print(f"First value: {record[0]}")
    print(f"Second value: {record[1]}")

    # Get value with default
    print(f"Email: {record.get('email', 'not provided')}")

    # Get single value
    print(f"Value of 'name': {record.value('name')}")
    print(f"Value at index 0: {record.value(0)}")
    print(f"Value with default: {record.value('missing', 'default')}")

    # Get multiple values
    print(f"All values: {record.values()}")
    print(f"Specific values: {record.values('name', 'age')}")

    # Get keys
    print(f"Keys: {record.keys()}")

    # Get items as key-value pairs
    print(f"Items: {list(record.items())}")

    # Convert to dictionary
    print(f"As dict: {record.data()}")
    print(f"Specific fields: {record.data('name', 'age')}")

    # Check if key exists
    print(f"Has 'name': {'name' in record}")

    # Iterate over record
    for key in record.keys():
        print(f"{key}: {record[key]}")

driver.close()
```

## Pandas DataFrame Integration

Convert results directly to pandas DataFrames for data analysis.

```python
from neo4j import GraphDatabase

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

with driver.session(database="neo4j") as session:
    # Setup test data
    session.run("MATCH (n) DETACH DELETE n").consume()
    session.run("""
        CREATE (a:Person {name: 'Alice', age: 30, joined: date('2020-01-15')})
        CREATE (b:Person {name: 'Bob', age: 25, joined: date('2021-06-20')})
        CREATE (c:Person {name: 'Charlie', age: 35, joined: date('2019-03-10')})
        CREATE (a)-[:KNOWS {since: 2020}]->(b)
        CREATE (b)-[:KNOWS {since: 2021}]->(c)
    """).consume()

    # Basic DataFrame conversion
    result = session.run("MATCH (p:Person) RETURN p.name AS name, p.age AS age")
    df = result.to_df()
    print(df)
    #       name  age
    # 0    Alice   30
    # 1      Bob   25
    # 2  Charlie   35

    # DataFrame with expanded graph types
    result = session.run("MATCH (p:Person) RETURN p")
    df = result.to_df(expand=True)
    print(df.columns.tolist())
    # ['p().prop.name', 'p().prop.age', 'p().prop.joined', 'p().element_id', 'p().labels']

    # Parse date/datetime columns
    result = session.run("MATCH (p:Person) RETURN p.name AS name, p.joined AS joined")
    df = result.to_df(parse_dates=True)
    print(df.dtypes)
    # joined column will be pandas Timestamp

driver.close()
```

## Bookmark Manager for Causal Consistency

Manage bookmarks across sessions to ensure causal consistency in a cluster.

```python
from neo4j import GraphDatabase

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

# Create a bookmark manager
bookmark_manager = GraphDatabase.bookmark_manager()

# All sessions using this manager will see each other's writes
with driver.session(database="neo4j", bookmark_manager=bookmark_manager) as session:
    session.run("CREATE (p:Person {name: 'Alice'})").consume()

# This session sees the write from above
with driver.session(database="neo4j", bookmark_manager=bookmark_manager) as session:
    result = session.run("MATCH (p:Person {name: 'Alice'}) RETURN p")
    record = result.single()  # Guaranteed to find Alice

# execute_query uses its own bookmark manager by default
driver.execute_query("CREATE (p:Person {name: 'Bob'})", database_="neo4j")

# Chain execute_query with session using the driver's bookmark manager
with driver.session(
    database="neo4j",
    bookmark_manager=driver.execute_query_bookmark_manager
) as session:
    # This session sees Bob created by execute_query
    result = session.run("MATCH (p:Person {name: 'Bob'}) RETURN p")
    record = result.single()

driver.close()
```

## Async Driver Usage

Use the async driver for asyncio-based applications.

```python
import asyncio
from neo4j import AsyncGraphDatabase, RoutingControl

async def main():
    driver = AsyncGraphDatabase.driver(
        "neo4j://localhost:7687",
        auth=("neo4j", "password")
    )

    # Verify connectivity
    await driver.verify_connectivity()

    # Simple query with execute_query
    records, summary, keys = await driver.execute_query(
        "CREATE (p:Person {name: $name}) RETURN p.name AS name",
        name="AsyncAlice",
        database_="neo4j"
    )
    print(f"Created: {records[0]['name']}")

    # Session-based operations
    async with driver.session(database="neo4j") as session:
        # Auto-commit query
        result = await session.run(
            "MATCH (p:Person) RETURN p.name AS name"
        )
        async for record in result:
            print(f"Found: {record['name']}")

        # Transaction function
        async def create_person(tx, name: str):
            result = await tx.run(
                "CREATE (p:Person {name: $name}) RETURN p.name AS name",
                name=name
            )
            record = await result.single()
            return record["name"]

        name = await session.execute_write(create_person, "AsyncBob")
        print(f"Created in tx: {name}")

    # Explicit transaction
    async with driver.session(database="neo4j") as session:
        async with await session.begin_transaction() as tx:
            await tx.run("CREATE (p:Person {name: 'AsyncCharlie'})")
            # Commits automatically

    await driver.close()

# Run the async code
asyncio.run(main())
```

## Error Handling

Handle various driver and database errors appropriately.

```python
from neo4j import GraphDatabase
from neo4j.exceptions import (
    Neo4jError,
    ClientError,
    CypherSyntaxError,
    ConstraintError,
    AuthError,
    ServiceUnavailable,
    SessionExpired,
    TransientError,
    DriverError,
    ConfigurationError,
    ResultConsumedError,
)

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

try:
    with driver.session(database="neo4j") as session:
        # Syntax error
        try:
            session.run("INVALID CYPHER SYNTAX").consume()
        except CypherSyntaxError as e:
            print(f"Syntax error: {e.message}")
            print(f"Error code: {e.code}")

        # Constraint violation
        try:
            session.run("CREATE CONSTRAINT FOR (p:Person) REQUIRE p.id IS UNIQUE").consume()
            session.run("CREATE (p:Person {id: 1})").consume()
            session.run("CREATE (p:Person {id: 1})").consume()  # Duplicate
        except ConstraintError as e:
            print(f"Constraint violation: {e.message}")

        # Accessing consumed result
        result = session.run("RETURN 1")
        result.consume()
        try:
            list(result)  # Result already consumed
        except ResultConsumedError as e:
            print(f"Result consumed: {e}")

except AuthError as e:
    print(f"Authentication failed: {e.message}")
except ServiceUnavailable as e:
    print(f"Database unavailable: {e}")
    print(f"Retryable: {e.is_retryable()}")
except TransientError as e:
    print(f"Transient error (retryable): {e.message}")
    print(f"Retryable: {e.is_retryable()}")
except Neo4jError as e:
    print(f"Database error: {e.message}")
    print(f"Code: {e.code}")
    print(f"GQL Status: {e.gql_status}")
except DriverError as e:
    print(f"Driver error: {e}")
finally:
    driver.close()
```

## Temporal Types

Work with Neo4j's temporal data types.

```python
from neo4j import GraphDatabase
from neo4j.time import Date, Time, DateTime, Duration
import pytz

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

with driver.session(database="neo4j") as session:
    # Create Date
    date = Date(2024, 1, 15)
    session.run("CREATE (e:Event {date: $date})", date=date).consume()

    # Create Time
    time = Time(14, 30, 0, nanosecond=500000000)  # 14:30:00.5
    session.run("CREATE (e:Event {time: $time})", time=time).consume()

    # Create DateTime with timezone
    tz = pytz.timezone('US/Eastern')
    dt = DateTime(2024, 1, 15, 14, 30, 0, tzinfo=tz)
    session.run("CREATE (e:Event {datetime: $datetime})", datetime=dt).consume()

    # Create Duration
    duration = Duration(months=1, days=5, seconds=3600)  # 1 month, 5 days, 1 hour
    session.run("CREATE (e:Event {duration: $duration})", duration=duration).consume()

    # Query temporal values
    result = session.run("""
        MATCH (e:Event)
        RETURN e.date AS date, e.time AS time, e.datetime AS datetime, e.duration AS duration
    """)
    for record in result:
        if record["date"]:
            d: Date = record["date"]
            print(f"Date: {d.year}-{d.month}-{d.day}")
        if record["time"]:
            t: Time = record["time"]
            print(f"Time: {t.hour}:{t.minute}:{t.second}")
        if record["datetime"]:
            dt: DateTime = record["datetime"]
            print(f"DateTime: {dt.iso_format()}")
        if record["duration"]:
            dur: Duration = record["duration"]
            print(f"Duration: {dur.months} months, {dur.days} days")

driver.close()
```

## Spatial Types

Work with Neo4j's spatial data types (Point).

```python
from neo4j import GraphDatabase
from neo4j.spatial import Point, CartesianPoint, WGS84Point

driver = GraphDatabase.driver("neo4j://localhost:7687", auth=("neo4j", "password"))

with driver.session(database="neo4j") as session:
    # Create 2D Cartesian point
    cart_point = CartesianPoint((1.5, 2.5))
    session.run("CREATE (l:Location {pos: $pos})", pos=cart_point).consume()

    # Create 3D Cartesian point
    cart_3d = CartesianPoint((1.0, 2.0, 3.0))
    session.run("CREATE (l:Location {pos3d: $pos})", pos=cart_3d).consume()

    # Create WGS84 point (latitude/longitude)
    geo_point = WGS84Point((-122.4194, 37.7749))  # San Francisco (lon, lat)
    session.run("CREATE (l:Location {geo: $geo})", geo=geo_point).consume()

    # Create WGS84 3D point (with altitude)
    geo_3d = WGS84Point((-122.4194, 37.7749, 100.0))  # With altitude
    session.run("CREATE (l:Location {geo3d: $geo})", geo=geo_3d).consume()

    # Query spatial values
    result = session.run("MATCH (l:Location) RETURN l.pos AS pos, l.geo AS geo")
    for record in result:
        if record["pos"]:
            p: Point = record["pos"]
            print(f"Cartesian: x={p.x}, y={p.y}, srid={p.srid}")
        if record["geo"]:
            g: Point = record["geo"]
            print(f"WGS84: lon={g.x}, lat={g.y}, srid={g.srid}")

    # Distance calculation using Cypher
    result = session.run("""
        WITH point({x: 0, y: 0}) AS p1, point({x: 3, y: 4}) AS p2
        RETURN point.distance(p1, p2) AS distance
    """)
    print(f"Distance: {result.single()['distance']}")

driver.close()
```

## Summary

The Neo4j Python Driver provides a comprehensive API for interacting with Neo4j databases from Python applications. The key usage patterns include: `driver.execute_query()` for simple, one-off queries with automatic session and transaction management; `session.run()` for auto-commit transactions; `session.execute_read()` and `session.execute_write()` for managed transaction functions with automatic retry; and `session.begin_transaction()` for explicit transaction control. The driver supports both synchronous and asynchronous programming models through separate `GraphDatabase` and `AsyncGraphDatabase` entry points.

For production applications, the recommended patterns are: use `execute_query()` for simple operations that don't require fine-grained control; use transaction functions (`execute_read`/`execute_write`) for operations that benefit from automatic retry on transient failures; use explicit transactions only when you need manual control over commit/rollback timing; always close drivers, sessions, and results using context managers (`with` statements); configure connection pools appropriately for your workload; and use bookmark managers when causal consistency across sessions is required.