Neo4j_interview.txt

# --------------------------------------------------------------
    # https://career.guru99.com/top-20-ne04j-interview-questions/
    # https://mindmajix.com/neo4j-interview-questions-answers
    # https://interviewprep.org/neo4j-interview-questions/
    # https://www.acte.in/neo4j-interview-questions-and-answers/
# --------------------------------------------------------------
1) What is Neo4j?
Neo4j is an open source NOSQL graph database, implemented in Java. 
It saves data structured in graphs rather than in tables.
# --------------------------------------------------------------
2) For what Neo4j is widely used for?
Neo4j is widely used for
 - Highly connected data - Social Network
 - Recommendations ( e-coomerce)
 - Path Finding
 - Data First Schema (bottom-up)
 - Schema Evolution
 - A* (Least Cost Path)
# --------------------------------------------------------------
3) How Neo4j differs from MySQL?
  - Neo4j consists of vertices and edges. 
  - Each vertex or node represent a key value or attribute.
  - It is possible to store dynamic content like images, videos, audio,
  - It has the capability for deep search into the database 
    without affecting the performance along with efficient timing
  - We can relate any two objects in neo4j by making relationship 
    between any two nodes
# --------------------------------------------------------------
4) What are some important characteristics of neo4j?
   Written in Java
   Continuous time traversals, 
   double linking between nodes and relationships.
   fast relationships 
   Memory caching, compact storage, 
# --------------------------------------------------------------
5) Explain Nodes, Relationships, Properties and Labels
   Nodes - entities
   Relationships - connect entities and structure domain
   Properties - meta-data and attributes
   Labels - group nodes by role
# --------------------------------------------------------------
6) Explain how you can run CQL commands in Neo4j?
   CQL = Cypher Query Language
   Can run from $ prompt
# --------------------------------------------------------------
7) Describe types of object caches in Neo4j
   There are two types: Reference Caches and High-performance Caches
# --------------------------------------------------------------
8) Describe query language
   Neo4j uses Cypher Query Language (CQL), which is unique to Neo4j.
        START n
        MATCH n-[r]- m
        RETURN r;
   Traversing the graph from (Start), 
   rules for traversal (Match) 
   and what to (Return)
# --------------------------------------------------------------
9) Is it possible to query Neo4j over the internet?
   Yes, you can query over the web, or run locally.
# --------------------------------------------------------------
10) How you create/delete databases in Neo4j ?
    You can do it form GUI (dropdown in Project)
    Also you can delete from terminal ( rm -rf data/* )
# --------------------------------------------------------------
11) Explain how Neo4j can be helpful in detecting Brute Force Attack?
    Neo4J allows to store and retrieve multiple complex relations
    in real time. For example, IP addresses, login failures, etc.
# --------------------------------------------------------------
12) How indexing is done in Neo4j?
    Automatic indexes can be created:
    START n=node:node_auto_index(name='abc') RETURN n
# --------------------------------------------------------------
13) Mention how files are stored in Neo4j?
    Neo4j stores graph data in store files, 
    for example, file for relationships, for nodes, for properties etc.:
       neostore.nodestore.db
       neostore.propertystore.db
       etc.
# --------------------------------------------------------------
14) List some CQL commands functionality
    Create nodes with and without properties
    Create a relationship between nodes (with/without properties)
    Make multiple or single labels to a Node or a Relationship
# --------------------------------------------------------------
15) Explain what Neo4j CQL MATCH command is used for?
    to get data about properties and nodes
    to get data about relationship, nodes and properties
# --------------------------------------------------------------
16) Explain the MATCH command syntax and how to use it.
    MATCH is used to define rules of traversing data in query
    Syntax:  MATCH ( <node-name>:<label-name> )
        START n
        MATCH n-[r]- m
        RETURN r;
# --------------------------------------------------------------
17) Explain what does the SET clause is used for in Neo4j?
    To update or add properties values(for nodes or relationships)
# --------------------------------------------------------------
18) Explain what Neo4j CQL LIMIT clause is used for?
    It is used to limit the number of returned rows in a query
# --------------------------------------------------------------
19) What is the "IN" Operator syntax in Neo4i?
    Example:  IN[ <Collection-of-values>]
# --------------------------------------------------------------
20) Explain how Neo4j stores primitive array?
    In a compressed form using a "bit saving" algorithm.
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
**) What is the local address?
    http://127.0.0.1:7474/
# --------------------------------------------------------------
4: List main features of Neo4J.
   - Data Representation using a graph model
   - Reliable ACID transactions.
   - Quick and Durable.
   - High-Speed query execution for traversals.
   - Easily accessible by REST or APIs.
# --------------------------------------------------------------
5: What does a Neo4J graph node store?
   Key-Value Pairs
# --------------------------------------------------------------
9: Give example of CQL query
   Example: To get cast of actors starting with S

    MATCH (actor:Person)-[:ACTED_IN]->(movie:Movie)
    WHERE movie.title STARTS WITH "S"
    RETURN movie.title AS title, collect(actor.name) AS cast
    ORDER BY title ASC LIMIT 10;
# --------------------------------------------------------------
10: Mention a few other famous graph databases available?
        Oracle NoSQL Database
        Sqrrl Enterprise
        Graph Story
        Titan
        Velocity Graph
        InfoGrid
        Bitsy
        Teradata Aster
        Oracle Spatial and Graph
        Info Grid
        etc.
# --------------------------------------------------------------
11: List out some of the Neo4J Commands you use?
        CREATE - To create a node or relationship.
        MATCH - To read or retrieve all the nodes in the database.
        MERGE - Combination of CREATE and MATCH.
        SET - To add or update properties to new or existing nodes/relationships.
        CREATE UNIQUE - To mention unique constraints in order to avoid redundant values.
# --------------------------------------------------------------
12: What is the REMOVE command used for?
    To remove labels and properties of the nodes
# --------------------------------------------------------------
13: What is the difference between REMOVE and DELETE commands?
    REMOVE is for labels and properties of nodes
    DELETE is to remove nodes and relationships. 
# --------------------------------------------------------------
14: What is object Cache in Neo4J?
    Object Cache improves performance by caching nodes and their properties
# --------------------------------------------------------------
15: What are the types of Object Cache in Neo4J?
    There are two types of Object Cache:
          Reference Cache
          High-Performance Cache(HPC)
# --------------------------------------------------------------
16: Which command is used to update properties or adding new properties to existing relationships?
    SET
# --------------------------------------------------------------
17: What is the difference between Neo4j Vs MongoDB?
    Neo4j is a Graph DB, written in Java, may have optional schema.
          uses CQL (Cypher Query Language) and APIs
          Uses triggers
# --------------------------------------------------------------
19: What is the IN Operator syntax?
    IN[ ]
# --------------------------------------------------------------
20: CREATE UNIQUE is used for?
    CREATE UNIQUE used for fixing the graph structures.
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
 - Neo4j supports ACID transactions
 - CQL - declarative graph query language
 - Nodes and relationships have direct pointers to each other, enabling fast traversal.
 - Index-free adjacency, meaning every node directly references its adjacent node, 
   eliminating the need for index lookups during traversals. This results in constant 
   time data retrieval regardless of the size of the data set.
 - high-availability setups through master-slave replication
 - sharding via graph partitioning
# --------------------------------------------------------------
3. How would you convert a relational database into a graph using Neo4j?
   Identify entities and relationships to represent as nodes and edges
   Use Neo4j ETL tool for importing data from your relational database. 
   This involves mapping relational data to graph data structures
   For complex transformations, Cypher's LOAD CSV command can be used
   Create indexes on frequently accessed nodes 
# --------------------------------------------------------------
4. What are the advantages of using Cypher Query Language (CQL) over SQL in Neo4j?
    - Cypher is designed for querying graph data natively
    - Cypher is declarative (like SQL)
    - CQL syntax visually represents the graph model
    - Cypher can perform deep traversals efficiently
# --------------------------------------------------------------
5. What is ACID compliance in the context of Neo4j and why is it important?
   ACID = Atomicity, Consistency, Isolation, and Durability
   - provides high data integrity, prevents data corruption, 
     enhances reliability & data security, ensures robustness
# --------------------------------------------------------------
7. How does indexing work in Neo4j and when should it be used?
   Indexes improve speed. There are native and schema indexes. 
   Native indexes are automatically maintained by Neo4j for all nodes and relationships
   Schema indexes need manual creation and are used for specific label-property combinations.
   Indexes = mapping between the indexed property values and the ids of the nodes or relationships.
   Indexes come at a cost of additional storage space and slower write operations due to index maintenance. 
# --------------------------------------------------------------
8. Can you elaborate on data replication and data partitioning strategies in Neo4j?
   Replication strategy “leader-follower” model. 
   Data partitioning in Neo4j is achieved manually by creating 
   multiple databases or using application-level sharding. 
   This involves splitting your data across different Neo4j instances 
   based on certain criteria like geographical location or user groups.
   This splitting introduces complexity in managing transactions 
   and maintaining consistency across shards. 
# --------------------------------------------------------------
9. How do you handle complex joins in Neo4j?
   There are relationships, not joins.
   For example, if you 2 types of nodes: Person and Company,
   then you can have relationship WORKS_FOR between them.
        MATCH (p:Person)-[:WORKS_FOR]->(c:Company {name:'Company Name'})
        RETURN p.name
# --------------------------------------------------------------
10. How would you model hierarchical data in Neo4j?
    Neo4j is ideal for modeling hierarchical data.
    Example of organizational chart:
        CREATE (e1:Employee {name:'John', position:'CEO'})
        CREATE (e2:Employee {name:'Jane', position:'Manager'})
        CREATE (e3:Employee {name:'Doe', position:'Developer'})
        CREATE
        (e1)-[:MANAGES]->(e2),
        (e2)-[:REPORTS_TO]->(e1),
        (e2)-[:MANAGES]->(e3),
        (e3)-[:REPORTS_TO]->(e2)
# --------------------------------------------------------------
11. Can you explain the difference between depth-first and breadth-first traversal in Neo4j?
        // Depth-first 
        MATCH p=(n)-[*]->(m)
        WHERE n.name = 'StartNode'
        RETURN p

        // Breadth-first
        MATCH p=(n)-[*0..]->(m)
        WHERE n.name = 'StartNode'
        RETURN p

        // You can adjust the depth of the traversal
        //     [*0..1] will only traverse one level deep, 
        //     [*0..3] will traverse up to three levels deep.
# --------------------------------------------------------------
12. Please describe the use of aggregates and collections in Cypher.
        Aggregates: COUNT(), AVG(), SUM(), MIN() and MAX()
                // count all nodes in database
                MATCH (n)
                RETURN COUNT(n) AS total_nodes;
                // count number of all relationships
                MATCH ()-[r]->()
                RETURN COUNT(r) AS total_relationships;
        Collections = ordered lists of elements. They can be created 
        using square brackets [], and manipulated with operators 
        such as IN (membership), + (concatenation), and * (repetition). 
        Collections also support list comprehension for filtering and transforming their elements.
# --------------------------------------------------------------
14. What is the role of labels in Neo4j and how do they aid in performance optimization?
    Labels in Neo4j are identifiers used to group nodes with similar characteristics. 
    When a label is assigned to a node, an index is automatically created.
    Labels also enable schema descriptions on parts of the graph and can be used to impose constraints.
# --------------------------------------------------------------
15. If you needed to perform full-text searching in Neo4j, how would you go about it?
    Create a full-text index on the nodes or relationships that you want to search
    CALL db.index.fulltext.createNodeIndex or createRelationshipIndex procedures
    Perform searches using db.index.fulltext.queryNodes or  queryRelationships procedures
    These return nodes or relationships whose specified properties match your query string.
    The query language used is Lucene's default operator set, allowing for complex queries 
    including wildcard, fuzzy, proximity, and range searches.
# --------------------------------------------------------------
16. Provide an example of using Neo4j for social network analysis.
    Example: nodes represent people and relationships represent friendships. 
        CREATE (Person1:Person {name:'John'}), 
            (Person2:Person {name:'Jane'}),
            (Person3:Person {name:'Doe'}),
            (Person1)-[:FRIEND ...
# --------------------------------------------------------------
17. Use cases for stored procedures ?
    Batch processing, data import/export, graph algorithms, network analysis. 
    Stored procedures also allow integration with other systems through APIs, 
    enabling the creation of custom functions.
# --------------------------------------------------------------
18. What is "Graph Density" ?
    Graph Density = Number of connections / Number of possible connections (N squared)
    More connections - slower performance.
    Mitigate by using vertical partitioning (split data into smaller groups) and/or indexing.
# --------------------------------------------------------------
19. How do you handle constraints and validations in Neo4j?
    Constraints (for nodes or relationships) are used to ensure data integrity.
    Examples: uniqueness, existence, and node key constraints. 
    Validations can be implemented using triggers or stored procedures. 
         CREATE CONSTRAINT ON (book:Book) ASSERT book.isbn IS UNIQUE
# --------------------------------------------------------------
20. What is "hot spots" ?
    Certain nodes or relationships are accessed very often causing performance issues.
    Mitigation: data partitioning accross several servers,
                caching in memory 
                indexing
                avoiding full scans
# --------------------------------------------------------------
21. What is the difference between embedded and server mode in Neo4j?
    Embedded mode - Neo4j runs within the application's JVM
    Server mode - runs as server, clients connect to it.
# --------------------------------------------------------------
22. How does Neo4j handle transactions and what makes it unique?
    ACID is supported (Atomicity, Consistency, Isolation, Durability).
    As all data relationships are stored directly in the DB itself,
    it is easier and faster to do ACID transactions.
# --------------------------------------------------------------
23. How would you use the Neo4j's REST API for CRUD operations?
    Create - POST request to /db/data/node with an optional JSON body for properties. 
    Read -   GET /db/data/node/{nodeId} to retrieve the node. 
    Update - PUT or POST to /db/data/node/{nodeId}/properties sending JSON with new property values
    Delete - DELETE /db/data/node/{nodeId} removes the node. 

    Relationships are similar: 
    Create - POST to /db/data/node/{startNodeId}/relationships creates one, specifying endNodeId and type in the body. 
    Read - GET /db/data/relationship/{relId} reads it
    Update - PUT or POST to /db/data/relationship/{relId}/properties updates
    Delete - DELETE /db/data/relationship/{relId} deletes.
# --------------------------------------------------------------
24. What are some common performance issues in Neo4j and how to mitigate?
    Inefficient queries  - avoid Cartesian products, reduce the number of returned nodes, 
       use 'PROFILE' or 'EXPLAIN' to understand how a query is executed.
    Improper indexing  - create indexes for frequently accessed properties. 
       Also, consider composite indexes if multiple properties are often queried together.
    Hardware limitations - increase RAM to avoid disk swapping.
# --------------------------------------------------------------
25. How do you ensure the security of data in Neo4j? 
    How to set up RBAC (role-based access control)?
    Neo4j provides authentication, authorization, and encryption. 
    For RBAC: 
      Create roles using the ROLE command. 
      Assign privileges to these roles, for example: 
           READ {*} ON GRAPH * TO role_name.
      Assign users to roles using the  ROLE command, for example:
           ROLE role_name TO user_name
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
# --------------------------------------------------------------
Example of CQL query:

    MATCH (n:Person)-[:FRIEND]->(friend)
    WHERE n.name = ‘John'
    RETURN n, friend
# --------------------------------------------------------------
10. What are some other well-known graph databases that are out there?
    Amazon Neptune
    Microsoft Azure Cosmos DB
    JanusGraph
    ArangoDB
    OrientDB
    AllegroGraph
# --------------------------------------------------------------
Which Neo4J commands do you frequently use?

    MATCH: Used to specify patterns to be matched in the graph.
    MATCH (n:Label)-[:RELATIONSHIP]->(m)
    CREATE: Used to create nodes and relationships in the graph.
    CREATE (n:Label {property: value})-[:RELATIONSHIP]->(m:Label)
    RETURN: Specifies what data to retrieve from a query.
    RETURN n.property, m.property
    WHERE: Applies a condition to filter results.
    WHERE n.property = ‘value'
    MERGE: Matches a pattern, or creates it if it doesn't exist.
    MERGE (n:Label {property: ‘value'})
# --------------------------------------------------------------
Which kinds of object caches are available in Neo4J?
   Node Cache & Relationship Cache
# --------------------------------------------------------------
How to add or update property?
        MATCH (a)-[r]->(b)
        SET r.newProperty = ‘NewValue'
# --------------------------------------------------------------
Example of using LIMIT clause
        MATCH (n)
        RETURN n
        LIMIT 10
# --------------------------------------------------------------
Example of using IN Operator?
        MATCH (n)
        WHERE n.propertyName IN ['value1', 'value2']
        RETURN n
# --------------------------------------------------------------
What is the purpose of CREATE UNIQUE?
    Creates only if it doesn't exist.
# --------------------------------------------------------------
Explain how collections and aggregates are used in Cypher?
   Collections like lists and aggregates like `COLLECT` 
   are used to manipulate and aggregate data. 
   For example, `COLLECT `can group results into a list, 
   while `UNWIND` can be used to transform a list into individual rows 
   for further processing.
# --------------------------------------------------------------
How can labels help with performance optimization?
    Limiting query by label reduces the scope of operations and improves 
    the overall speed of graph traversals and searches
# --------------------------------------------------------------
How would you approach full-text searching in Neo4j if you needed to?
   third-party plugins like the Neo4j Elasticsearch Integration 
   or the `fulltext` index offered by the APOC library 
# --------------------------------------------------------------
Social network analysis examples:
    MATCH (u)-[r]-(v) RETURN u, count(r) AS degree ORDER BY degree DESC LIMIT 10
# --------------------------------------------------------------
Does Neo4J store files in nodes?
    No, usually only path to files.
# --------------------------------------------------------------
What is a relationship 
    relationships are node-to-node connections
    They have type, optional characteristics, and direction
# --------------------------------------------------------------
Neo4j supports integration with Java, Python, and JavaScript.
# --------------------------------------------------------------
How is scalability handled by Neo4j?
    Clustering and sharding
# --------------------------------------------------------------
In Neo4j you can have nodes without any connections:
      CREATE (n:Person {name: 'John Doe'})

You can add connections:
      MATCH (n:Person {name: 'John Doe'}), (m:Person {name: 'Lisa Smith'})
      CREATE (n)-[:KNOWS]->(m)
# --------------------------------------------------------------
In Neo4j, nodes belong to a single graph by default. 
It is possible to create separate graphs (subgraphs) within a single database.
It is possible to create separate databases within the same instance. 
# --------------------------------------------------------------
What is label?
    A node may have 0, 1, or more lables.
    Labels use to narrow down the query to make them faster
# --------------------------------------------------------------
Whaat os Index Provider ?
    Index Provider determines how indexes are stored and queried.
# --------------------------------------------------------------
What is the Neo4j Bolt protocol ?
    Bolt is a binary protocol to effectively communicate with Noe4j server.
# --------------------------------------------------------------
Neo4j Import Tool 
    can import massive amounts of data from external sources in CSV and JSON formats
# --------------------------------------------------------------
Neo4j primarily employs an optimistic locking strategy for handling concurrency.
Each node has a version property, and transactions check this version before committing changes to identify conflicts. 
The explicit pessimistic locking is not a core feature.
# --------------------------------------------------------------
What is Neo4j Aura platform ?
    It is Neo4j as a managed cloud service.
# --------------------------------------------------------------
How is geographical data supported by Neo4j?
    Mapping and GIS analysis is handled using built-in indexing and spatial types
# --------------------------------------------------------------
Triggers - there is no built-in triggers
# --------------------------------------------------------------
What is Neo4j Extension ?
    a specially designed plugin or module that incorporates additional 
    custom processes, methods, or endpoints. Can be written in Java, Python, or Javascript.
# --------------------------------------------------------------
How are data backup and recovery handled by Neo4j?
    Neo4j has both online and offline backup techniques. 
    Recovery from data loss includes replaying transaction logs, 
    applying incremental backups as necessary, and recovering 
    the database from a backup.
# --------------------------------------------------------------
What is Neo4j Sandbox ?
    It is a cloud-based pre-configured platform to explore and learn Neo4j
# --------------------------------------------------------------
How time and date can be stored/used?
   Date/time attributes on nodes or relationships may be used.
# --------------------------------------------------------------
What is Neo4j Graph Data Science library?
    Graph analytics, Graph Embeddings, Graph Projection
# --------------------------------------------------------------
There is no explicit COMMIT and ROLLBACK statements. 
Instead each Cypher query is automatically wrapped into a transaction.
# --------------------------------------------------------------
In a distributed context, Neo4j uses the Raft consensus algorithm 
to guarantee data consistency. 
# --------------------------------------------------------------
Neo4j uses graph processing engines. Supports parallel operation,
leverages multi-core operations and multiple computers.
Leverages caching and indexing.
# --------------------------------------------------------------
Neo4j has Connectors and Drivers (Java, Python, .NET, ...) 
RESTful API over HTTP
ETL Tools like Apache Kafka, NiFi, or Talend
# --------------------------------------------------------------
PageRank algorithm - to rank nodes
    // Install Graph Data Science library
    CALL gds.alpha.enable
    // Run query to retrieve PageRank scores
        CALL gds.pageRank.write({
        nodeProjection: '*',
        relationshipProjection: {
            REL: {
            type: 'REL',
            direction: 'OUTGOING'
            }
        },
        relationshipWeightProperty: 'weight',
        writeProperty: 'rank'
        })
        YIELD nodePropertiesWritten, relationshipPropertiesWritten, commitTime, writeMillis

        MATCH (n)
        RETURN n.name, n.rank
        ORDER BY n.rank DESC
        LIMIT 10
# --------------------------------------------------------------
Finding clusters / communities
    Neo4j's has algorithms for community detection, such as Louvain and Label Propagation.
    They unveil organic clusters, unveiling significant connections and configurations.
    This clustering can be used for Analysis and Recommendations
# --------------------------------------------------------------
Date and time:
    LocalDate, LocalTime, DateTime
    date(), time(), datetime() functions
    Temporal indexing, Temporal Range Queries (FROM, TO, DURING) 
# --------------------------------------------------------------
Importing data from CSV file

        // Create the nodes for the users
        USING PERIODIC COMMIT 1000
        LOAD CSV WITH HEADERS FROM 'file:///users.csv' AS row
        MERGE (u:User {id: row.id})
        ON CREATE SET u.name = row.name, u.age = row.age

        // Create the nodes for the posts
        USING PERIODIC COMMIT 1000
        LOAD CSV WITH HEADERS FROM 'file:///posts.csv' AS row
        MERGE (p:Post {id: row.id})
        ON CREATE SET p.title = row.title, p.content = row.content

        // Create the relationships between users and posts
        USING PERIODIC COMMIT 1000
        LOAD CSV WITH HEADERS FROM 'file:///user_posts.csv' AS row
        MATCH (u:User {id: row.user_id}), (p:Post {id: row.post_id})
        MERGE (u)-[:AUTHOR]->(p)

# --------------------------------------------------------------
Neo4j for social network analysis.
Neo4j for recommendation systems?
Neo4j for Anomaly Detection
# --------------------------------------------------------------
Example of User-defined functions (UDF) creation and usage
    CALL apoc.help("apoc")
    CALL apoc.trigger.add("absolute_value", "RETURN Math.abs(value)", 
        { value: { type: "number", description: "The number to get the absolute value of" } })

    CALL apoc.trigger.run("absolute_value", { value: -4 }) 
    YIELD value
    RETURN value

Note: "APOC" stands for "Awesome Procedures On Cypher."
# -------------------------------------------------------------- 
Importing lots of data using the neo4j-admin import tool
    Importing Data Offline (uses parallelism)
# -------------------------------------------------------------- 
Connected Components algorithm - identify clusters
# -------------------------------------------------------------- 
Shortest Path algorithm in Neo4j's Graph Algorithms library
    gds.alpha.shortestPath
# -------------------------------------------------------------- 
Betweenness Centrality algorithm
    figuring out the percentage of shortest paths that go through each node, 
    highlighting important bridges or connectors in the graph 
    that are essential to network resilience or information flow
# -------------------------------------------------------------- 
Label Propagation algorithm - detect clusters
# -------------------------------------------------------------- 
spatial indexing, location-based queries (by geographic coordinates)
# -------------------------------------------------------------- 
Neo4j Bloom tool - interactive visual exploration
# -------------------------------------------------------------- 
undirected vs directed relationship
    CREATE (p)-[:LIKES]-(m)       // undirected
    CREATE (p)-[:WATCHED]->(m)    // directed
# -------------------------------------------------------------- 
Removal of the node - consider cascading effects on connected nodes
        MATCH (u:User)
        WHERE u.age > 30
        DETACH DELETE u
# --------------------------------------------------------------