# CSV Graph Loader Examples Complete loader examples for different domains and use cases. ## Example 1: Business Domain - Employees & Companies ### Input CSV ```csv employee_id,employee_name,job_title,department,company_name,company_industry,manager_id E001,Alice Johnson,Software Engineer,Engineering,Acme Corp,Technology,E100 E002,Bob Smith,Product Manager,Product,Acme Corp,Technology,E100 E003,Carol Davis,Sales Executive,Sales,Acme Corp,Technology,NULL E004,David Wilson,Software Engineer,Engineering,TechStart Inc,Technology,E200 E005,Emma Brown,HR Manager,HR,Acme Corp,Technology,E100 ``` ### Entity Detection ``` Entities detected: - Person: employee_id, employee_name, job_title - Company: company_name, company_industry - Department: department ``` ### Generated Neo4j Cypher ```cypher LOAD CSV WITH HEADERS FROM 'file:///employees.csv' AS row -- Create Person nodes MERGE (e:Person {id: row.employee_id}) SET e.name = row.employee_name, e.jobTitle = row.job_title, e.department = row.department -- Create Company nodes MERGE (c:Company {name: row.company_name}) SET c.industry = row.company_industry -- Create WORKS_AT relationships MERGE (e)-[:WORKS_AT]->(c) -- Create REPORTS_TO relationships (if manager exists) WITH row, e WHERE row.manager_id IS NOT NULL MATCH (manager:Person {id: row.manager_id}) MERGE (e)-[:REPORTS_TO]->(manager) ``` ### Generated RDF Triples (Turtle) ```turtle @prefix ex: . @prefix foaf: . ex:person_E001 a ex:Person ; foaf:name "Alice Johnson" ; ex:jobTitle "Software Engineer" ; ex:department "Engineering" ; ex:worksAt ex:acme_corp ; ex:reportsTo ex:person_E100 . ex:person_E100 a ex:Person ; foaf:name "Manager Name" ; ex:jobTitle "Engineering Manager" . ex:acme_corp a ex:Company ; foaf:name "Acme Corp" ; ex:industry "Technology" . ``` ### Generated Property Graph JSON ```json { "nodes": [ {"id": "e1", "type": "Person", "properties": {"name": "Alice Johnson", "job_title": "Software Engineer", "department": "Engineering"}}, {"id": "e2", "type": "Person", "properties": {"name": "Bob Smith", "job_title": "Product Manager", "department": "Product"}}, {"id": "c1", "type": "Company", "properties": {"name": "Acme Corp", "industry": "Technology"}}, {"id": "c2", "type": "Company", "properties": {"name": "TechStart Inc", "industry": "Technology"}} ], "edges": [ {"source": "e1", "target": "c1", "type": "WORKS_AT"}, {"source": "e2", "target": "c1", "type": "WORKS_AT"}, {"source": "e1", "target": "e100", "type": "REPORTS_TO"}, {"source": "e2", "target": "e100", "type": "REPORTS_TO"} ], "metadata": { "total_persons": 5, "total_companies": 2, "total_relationships": 7 } } ``` --- ## Example 2: Scientific Domain - Research Papers ### Input CSV ```csv paper_id,title,publication_year,doi,author_name,author_affiliation,research_area P001,Deep Learning Advances,2023,10.1234/paper1,Dr. Alice Chen,MIT,Machine Learning P001,Deep Learning Advances,2023,10.1234/paper1,Dr. Bob Kumar,Stanford,Machine Learning P002,Quantum Computing,2023,10.1234/paper2,Dr. Carol Wilson,IBM,Quantum Computing P003,Gene Therapy,2024,10.1234/paper3,Dr. David Lee,Johns Hopkins,Biotechnology ``` ### Entity Detection & Deduplication ``` Entities detected: - Paper: paper_id, title, publication_year, doi - Researcher: author_name, author_affiliation - ResearchArea: research_area Deduplication: - Paper P001 appears twice (different authors) - Merged into single Paper node with multiple AUTHORS relationships ``` ### Generated Neo4j Cypher ```cypher LOAD CSV WITH HEADERS FROM 'file:///papers.csv' AS row -- Create unique Paper nodes MERGE (p:Paper {doi: row.doi}) SET p.id = row.paper_id, p.title = row.title, p.publication_year = toInteger(row.publication_year) -- Create Researcher nodes MERGE (r:Researcher {name: row.author_name}) SET r.affiliation = row.author_affiliation -- Create ResearchArea nodes MERGE (a:ResearchArea {name: row.research_area}) -- Create relationships MERGE (r)-[:AUTHORED]->(p) MERGE (p)-[:FOCUSES_ON]->(a) ``` ### Generated RDF/OWL ```turtle @prefix ex: . @prefix dcterms: . ex:paper_P001 a ex:ResearchPaper ; dcterms:title "Deep Learning Advances" ; dcterms:issued "2023"^^xsd:gYear ; ex:doi "10.1234/paper1" ; ex:hasAuthor ex:researcher_alice_chen, ex:researcher_bob_kumar ; ex:focusesOn ex:area_machine_learning . ex:researcher_alice_chen a ex:Researcher ; foaf:name "Dr. Alice Chen" ; ex:affiliation "MIT" . ex:area_machine_learning a ex:ResearchArea ; rdfs:label "Machine Learning" . ``` --- ## Example 3: Social Network Domain ### Input CSV ```csv user_id,username,display_name,bio,follower_count,account_type,followed_user_id,followed_username,post_id,post_text,post_date,engagement_count U001,alice_tech,Alice,Data Scientist,5000,verified,U002,bob_dev,POST1,Great insights on graphs,2024-01-15,250 U001,alice_tech,Alice,Data Scientist,5000,verified,U003,carol_ai,POST1,Great insights on graphs,2024-01-15,250 U002,bob_dev,Bob,Software Dev,3000,standard,U001,alice_tech,POST2,Working on ML,2024-01-16,180 ``` ### Entity Detection ``` Entities detected: - User: user_id, username, display_name, bio - Post: post_id, post_text, post_date - Hashtag/Topic: (implicit from post content) Relationships: - FOLLOWS: User → User - POSTED: User → Post - HAS_ENGAGEMENT: Post → Post/User ``` ### Generated Neo4j Cypher ```cypher LOAD CSV WITH HEADERS FROM 'file:///social.csv' AS row -- Create User nodes MERGE (u:User {id: row.user_id}) SET u.username = row.username, u.displayName = row.display_name, u.bio = row.bio, u.followerCount = toInteger(row.follower_count), u.accountType = row.account_type -- Create followed User nodes MERGE (following:User {id: row.followed_user_id}) SET following.username = row.followed_username -- Create Post nodes MERGE (p:Post {id: row.post_id}) SET p.text = row.post_text, p.postDate = datetime(row.post_date), p.engagementCount = toInteger(row.engagement_count) -- Create relationships MERGE (u)-[:FOLLOWS]->(following) MERGE (u)-[:POSTED]->(p) ``` ### Generated Property Graph JSON ```json { "nodes": [ {"id": "u1", "type": "User", "properties": {"username": "alice_tech", "display_name": "Alice", "account_type": "verified"}}, {"id": "u2", "type": "User", "properties": {"username": "bob_dev", "display_name": "Bob", "account_type": "standard"}}, {"id": "post1", "type": "Post", "properties": {"text": "Great insights on graphs", "engagement_count": 250}} ], "edges": [ {"source": "u1", "target": "u2", "type": "FOLLOWS"}, {"source": "u1", "target": "post1", "type": "POSTED"} ] } ``` --- ## Example 4: E-Commerce Domain ### Input CSV ```csv product_id,product_name,category,price,stock,seller_id,seller_name,order_id,customer_id,order_date,quantity,review_score PROD001,Laptop,Electronics,999.99,50,SELLER001,TechHub,ORD001,CUST001,2024-01-10,1,4.5 PROD002,Mouse,Electronics,29.99,200,SELLER001,TechHub,ORD002,CUST002,2024-01-11,2,4.0 PROD001,Laptop,Electronics,999.99,50,SELLER001,TechHub,ORD003,CUST001,2024-01-12,1,4.5 PROD003,Keyboard,Electronics,79.99,150,SELLER002,ElectroStore,ORD004,CUST003,2024-01-13,1,4.8 ``` ### Deduplication Strategy ``` Input rows: 4 Unique products: 3 (PROD001 appears twice) Unique orders: 4 Unique customers: 3 Unique sellers: 2 Merge strategy: - Product with same ID merged - Each order is unique even with same product ``` ### Generated Node & Edge Tables **nodes.csv** ```csv id,type,name,category,stock,price PROD001,Product,Laptop,Electronics,50,999.99 PROD002,Product,Mouse,Electronics,200,29.99 PROD003,Product,Keyboard,Electronics,150,79.99 SELLER001,Seller,TechHub,, SELLER002,Seller,ElectroStore,, CUST001,Customer,Customer 1,, CUST002,Customer,Customer 2,, ``` **edges.csv** ```csv source,target,type,quantity,review_score CUST001,PROD001,PURCHASED,1,4.5 CUST002,PROD002,PURCHASED,2,4.0 CUST001,PROD001,PURCHASED,1,4.5 CUST003,PROD003,PURCHASED,1,4.8 SELLER001,PROD001,SELLS,, SELLER001,PROD002,SELLS,, SELLER002,PROD003,SELLS,, ``` --- ## Example 5: Knowledge Graph - Wikipedia-like Data ### Input CSV ```csv entity_id,entity_name,entity_type,birth_date,profession,country,related_entity,relation_type E001,Marie Curie,Person,1867-11-24,Scientist,Poland,E002,MARRIED_TO E001,Marie Curie,Person,1867-11-24,Scientist,Poland,E003,DISCOVERED E002,Pierre Curie,Person,1859-05-15,Scientist,France,E001,MARRIED_TO E003,Radium,Chemical Element,,Element,Natural,E001,DISCOVERED_BY E004,Polonium,Chemical Element,,Element,Natural,E001,DISCOVERED_BY ``` ### Generated RDF Knowledge Graph ```turtle @prefix ex: . @prefix rdf: . @prefix rdfs: . @prefix foaf: . @prefix dbo: . ex:marie_curie a ex:Person ; foaf:name "Marie Curie" ; dbo:birthDate "1867-11-24"^^xsd:date ; dbo:profession "Scientist" ; dbo:country ex:poland ; ex:marriedTo ex:pierre_curie ; ex:discovered ex:radium, ex:polonium ; rdfs:comment "Polish-born physicist" . ex:radium a ex:ChemicalElement ; rdfs:label "Radium" ; ex:discoveredBy ex:marie_curie ; ex:country ex:natural ; ex:atomicNumber 88 . ex:pierre_curie a ex:Person ; foaf:name "Pierre Curie" ; dbo:birthDate "1859-05-15"^^xsd:date ; dbo:profession "Scientist" ; dbo:country ex:france ; ex:marriedTo ex:marie_curie . ex:poland a ex:Country ; rdfs:label "Poland" . ``` --- ## Loader Comparison Table | Domain | CSV Rows | Entities | Relationships | Node Types | Edge Types | Deduplication | Format | |--------|----------|----------|---------------|-----------|-----------|----------------|--------| | Business | 5 | 3 types | Manager+Company | Person, Company, Dept | WORKS_AT, REPORTS_TO | Merge by ID | Cypher | | Scientific | 4 | 3 types | Author+Area | Paper, Researcher, Area | AUTHORED, FOCUSES_ON | Merge duplicates | RDF/Turtle | | Social | 3 | 2 types | Follow+Post | User, Post | FOLLOWS, POSTED | Deduplicate | JSON | | E-Commerce | 4 | 3 types | Purchase+Sell | Product, Customer, Seller | PURCHASED, SELLS | Merge products | CSV tables | | Knowledge | 5 | 2 types | Multiple | Person, Element | MARRIED_TO, DISCOVERED | None | RDF/OWL | --- ## Transformation Performance Notes | Example | Input Size | Processing Time | Nodes Generated | Edges Generated | |---------|-----------|-----------------|-----------------|-----------------| | Business | 5 rows | < 100ms | 8 | 7 | | Scientific | 4 rows | < 100ms | 6 | 8 | | Social | 3 rows | < 50ms | 4 | 3 | | E-Commerce | 4 rows | < 100ms | 8 | 8 | | Knowledge | 5 rows | < 100ms | 6 | 6 | --- See [loader-patterns.md](../references/loader-patterns.md) for detailed CSV loader design patterns.