Neo4j
- Console Neo4j
- Using Neo4j from Python
- Neo4j Cypher Refcard
- Books
- Introducing the Neo4j 7-Day Sandbox
$ brew install neo4j/usr/local/Cellar/neo4j/2.3.2/libexec/conf
例如要能遠端連進 neo4j, 可以修改 neo4j-server.properties
$ neo4j start預設會運行在 port 7474, 第一次安裝會要求設定密碼, 使用者名稱: neo4j
org.neo4j.server.webserver.address=0.0.0.0
在 Neo4j, query 的語法叫做 Cypher
- MATCH: 像 SELECT
- CREATE: 新增 DB, 或新增 relationship
- WHERE: 條件
- RETURN: 回傳對象
語法起源來自 Ascii art
$ MATCH (user)-[:BIRTH_PLACE]->(country) RETURN user, country;
產生 Employee 的節點
- n: variable
- Employee: label
- {}: properties
$ CREATE (n:Employee {Name:"Chang-Wu", Surname: "Chen", Gender: "M"});
$ MATCH (n:Employee) RETURN n;
$ MATCH (n:Employee) RETURN n LIMIT 100;
$ MATCH (n:Employee) RETURN n.Gender;
注意 Neo4j 在回傳超過 1000 筆資料時, 會有錯誤, 所以最好的方式是在查詢資料返回時, 加上筆數限制
$ MATCH (n:Employee) RETURN n.Name AS Name, n.Gender AS Gender;
取回資料可用 AS 來重新命名欄位名稱
Neo4j web console 可用來查看 query, 或下指令, 亦可匯出 .json 或 .csv 檔案, 但是無法用來匯入資料, 操作資料透過 command line , 或是利用其他語言 api 來操作 Neo4j 比較好, 另外要繪製圖形需要自己處理, 可以使用其他套件, 例如 d3.js
修改資料可用 SET
$ MATCH (n:Employee {Name:"Chang-Wu"}) SET n.Gender=NULL;
刪除 Name 與 Gender
$ MATCH (n:Employee {Surname:"Chen"}) SET n.Name=NULL RETURN n;
刪除節點
$ MATCH (n:Employee {Surname:"Chen"}) DELETE n;
查看資料庫中所有資料
$ MATCH n RETURN n;
如果要刪除的點, 與其他點之間有關係的話, 必須先移除關係, 才能移除節點.
$ MATCH (n:Employee) WHERE n.Name="Chang-Wu" RETURN n;
$ MATCH (n:Employee) WHERE n.Name=~"Cha.*" RETURN n;
$ MATCH (n:Employee) WHERE n.Name STARTS WITH "Ch" RETURN n;
在 Neo4j 沒有所謂的 primary key, 換言之, 資料回重複出現兩次
$ CREATE (n:Employee {Name:"Chang-Wu", Surname: "Chen", Gender: "M"});
$ CREATE (n:Employee {Name:"Chang-Wu", Surname: "Chen", Gender: "M"});
Data
$ MATCH n RETURN n;
{
"results": [
{
"columns": [
"n"
],
"data": [
{
"row": [
{
"Gender": "M",
"Surname": "Chen",
"Name": "Chang-Wu"
}
],
"graph": {
"nodes": [
{
"id": "2",
"labels": [
"Employee"
],
"properties": {
"Gender": "M",
"Surname": "Chen",
"Name": "Chang-Wu"
}
}
],
"relationships": []
}
},
{
"row": [
{
"Gender": "M",
"Surname": "Chen",
"Name": "Chang-Wu"
}
],
"graph": {
"nodes": [
{
"id": "3",
"labels": [
"Employee"
],
"properties": {
"Gender": "M",
"Surname": "Chen",
"Name": "Chang-Wu"
}
}
],
"relationships": []
}
}
],
"stats": {
"contains_updates": false,
"nodes_created": 0,
"nodes_deleted": 0,
"properties_set": 0,
"relationships_created": 0,
"relationship_deleted": 0,
"labels_added": 0,
"labels_removed": 0,
"indexes_added": 0,
"indexes_removed": 0,
"constraints_added": 0,
"constraints_removed": 0
}
}
],
"errors": []
}但 Neo4j 中, 在生成 node 時, 會有不同的 node id 區別同樣的 data
$ CREATE (e:Employee {Name: "Chang-Wu"})-[:WORK_IN]->(c:Company {Name: "Yahoo!"});
$ CREATE (s:Skill {Name: "Neo4j"})<-[:KNOWNS]-(e:Employee {Name: "Chang-Wu"})-[:WORK_IN]->(c:Company {Name:"Yahoo!"});
$ MATCH (e:Employee {Name: "Chang-Wu"}), (s:Skill) WHERE s.Name="Neo4j" CREATE UNIQUE (e)-[:KNOWS]->(s);
CREATE UNIQUE用來避免資料庫重複新增同樣 data, 前面提過, Neo4j 沒有主鍵, 資料可被重複新增
找出關係
$ MATCH (e:Employee {Name: "Chang-Wu"})-[r]-(c:Company {Name: "Yahoo!"}) RETURN r;
$ MATCH (e:Employee)-[:KNOWS]->(s:Skill) RETURN e.Name, s.Name;
$ MATCH (e:Employee)-[:KNOWS]->(s:Skill) WHERE s.Name="Neo4j" RETURN e.Name;
$ MATCH (e:Employee)-[r]->(s:Skill) RETURN r;
刪除所有關係
$ start r=relationship(*) delete r;
from py2neo import Graph, Node, Relationship
graph = Graph()
a = Node("Person", name="Alice")
b = Node("Person", name="Bob")
ab = Relationship(a, "KNOWS", b, Since=2015)
graph.create(ab)
搜尋
a = graph.find_one("Person", "name", "Alice")
print a
nodes = graph.find("Person")
for node in nodes:
print node修改
a = graph.find_one("Person", "name", "Alice")
a.properties["name"] = "Anna"
a.push刪除
a = graph.find_one("Person", "name", "Alice")
a.delete()
graph.delete_all() # MATCH (n)-[r]-(q) DELETE n, r, q;執行 Cypher Query
print graph.cypher.execute("MATCH n RETURN n;")UNIQUE 另外也很重要, 當不想要重複新增資料