neo4j desktop 导入finbench数据集，使用命令导入

qq_46197858

已于 2024-03-21 14:10:51 修改

阅读量1k

点赞数 19

文章标签： neo4j

于 2024-03-20 10:01:59 首次发布

本文链接：https://blog.csdn.net/qq_46197858/article/details/136848349

版权

首先看一下整个数据集的组织结构

这里是数据集的下载地址

https://drive.google.com/drive/folders/1NIAo4KptskBytbXoOqmF3Sto4hTX3JIH

我选择的数据集是sf0.01,是最小的一个，选择的是其中snapshot里面的数据。

首先我是先打算导入节点信息，即 Loan,Account,Medium,Company,Person

顶点导入

首先是Loan

LOAD CSV WITH HEADERS FROM "file:///snapshot/Loan.csv" AS row
FIELDTERMINATOR '|'
MERGE (loan:Loan {Id: row.loanId})
ON CREATE SET loan.loanAmount = toFloat(row.loanAmount),
              loan.balance = toFloat(row.balance),
              loan.usage = row.loanUsage,
              loan.interestRate = toFloat(row.interestRate)

首先第一个问，因为文件中是用‘|’作为分隔符，所以要指定‘|’作为分隔符，即命令FIELDTERMINATOR '|'，

后面的几个顶点都是相同的操作

Person:

LOAD CSV WITH HEADERS FROM "file:///snapshot/Person.csv" AS row
FIELDTERMINATOR '|'
MERGE (person:Person {Id: row.personId})
ON CREATE SET person.name= row.personName,
              person.isBlocked= toBoolean(row.isBlocked),
              person.createTime= row.createTime,
              person.gender= row.gender,
              person.birthday=row.birthday,
              person.country= row.country,
              person.city= row.city

因为csv文件中，关于时间的标准，貌似并不是neo4j支持的ISO 8601标准，所以我并没有用datetime()函数去处理，而是直接默认他是string。

Medium

LOAD CSV WITH HEADERS FROM "file:///snapshot/Medium.csv" AS row
FIELDTERMINATOR '|'
MERGE (medium:Medium {Id: row.mediumId})
ON CREATE SET medium.type=row.mediumType,
              medium.isBlocked= toBoolean(row.isBlocked),
              medium.createTime= row.createTime,
              medium.lastLoginTime=row.lastLoginTime,
              medium.riskLevel=row.riskLevel

时间问题仍然是使用string类型代替

Account

LOAD CSV WITH HEADERS FROM "file:///snapshot/Account.csv" AS row
FIELDTERMINATOR '|'
MERGE (p:Account {Id: row.accountId})
ON CREATE SET p.createTime=row.createTime,
              p.isBlocked= toBoolean(row.isBlocked),
              p.type=row.accountType,
              p.nickname=row.nickname,
              p.phoneNumber=row.phonenum,
              p.email=row.email,
              p.freqLoginType=row.freqLoginType,
              p.lastLoginTime=row.lastLoginTime,
              p.accountLevel=row.accountLevel

Conpany

LOAD CSV WITH HEADERS FROM "file:///snapshot/Company.csv" AS row
FIELDTERMINATOR '|'
MERGE (p:Company {Id: row.companyId})
ON CREATE SET p.name=row.companyId,
              p.isBlocked= toBoolean(row.isBlocked),
              p.createTime=row.createTime,
              p.country=row.country,
              p.city=row.city,
              p.business=row.business,
              p.description=row.description,
              p.url=row.url

关系导入

repay关系（Account---->Loan）

LOAD CSV WITH HEADERS FROM "file:///snapshot/AccountRepayLoan.csv" AS row
FIELDTERMINATOR '|'
MATCH (account:Account {Id: row.accountId})
MATCH (loan:Loan {Id: row.loanId})
CREATE (account)-[:repay {amount: toFloat(row.amount), createTime: row.createTime}]->(loan)

transfer关系（account----->account）

LOAD CSV WITH HEADERS FROM "file:///snapshot/AccountTransferAccount.csv" AS row
FIELDTERMINATOR '|'
MATCH (account1:Account {Id: row.fromId})
MATCH (account2:Account {Id: row.toId})
CREATE (account1)-[:transfer {amount: toFloat(row.amount), createTime: row.createTime,orderNum:row.orderNum,comment:row.comment,payType:row.payType,goodsType:row.goodsType}]->(account2)

withdraw关系（account----->account）

LOAD CSV WITH HEADERS FROM "file:///snapshot/AccountTransferAccount.csv" AS row
FIELDTERMINATOR '|'
MATCH (account1:Account {Id: row.fromId})
MATCH (account2:Account {Id: row.toId})
CREATE (account1)-[:withraw {amount: toFloat(row.amount), createTime: row.createTime}]->(account2)

apply关系（company----->loan）

LOAD CSV WITH HEADERS FROM "file:///snapshot/CompanyApplyLoan.csv" AS row
FIELDTERMINATOR '|'
MATCH (company:Company {Id: row.companyId})
MATCH (loan:Loan {Id: row.loanId})
CREATE (company)-[:apply { createTime: row.createTime,org:row.org}]->(loan)

guarantee关系（company-------->company）

LOAD CSV WITH HEADERS FROM "file:///snapshot/CompanyGuaranteeCompany.csv" AS row
FIELDTERMINATOR '|'
MATCH (p1:Company {Id: row.fromId})
MATCH (p2:Company {Id: row.toId})
CREATE (p1)-[:guarantee { createTime: row.createTime,relation:row.relation}]->(p2)

invest关系（company----->company）

LOAD CSV WITH HEADERS FROM "file:///snapshot/CompanyInvestCompany.csv" AS row
FIELDTERMINATOR '|'
MATCH (p1:Company {Id: row.investorId})
MATCH (p2:Company {Id: row.companyId})
CREATE (p1)-[:invest { ratio:tofloat(row.ratio),createTime: row.createTime}]->(p2)

own关系（company------>account）

LOAD CSV WITH HEADERS FROM "file:///snapshot/CompanyOwnAccount.csv" AS row
FIELDTERMINATOR '|'
MATCH (p1:Company {Id: row.companyId})
MATCH (p2:Account {Id: row.accountId})
CREATE (p1)-[:own {createTime: row.createTime}]->(p2)

deposit关系（loan------>account）

LOAD CSV WITH HEADERS FROM "file:///snapshot/LoanDepositAccount.csv" AS row
FIELDTERMINATOR '|'
MATCH (p1:Loan {Id: row.loanId})
MATCH (p2:Account {Id: row.accountId})
CREATE (p1)-[:deposit {amount:tofloat(row.amount),createTime: row.createTime}]->(p2)

signln关系（medium------>account）

LOAD CSV WITH HEADERS FROM "file:///snapshot/MediumSignInAccount.csv" AS row
FIELDTERMINATOR '|'
MATCH (p1:Medium{Id: row.mediumId})
MATCH (p2:Account {Id: row.accountId})
CREATE (p1)-[:signin {createTime: row.createTime,location:row.location}]->(p2)

apply关系（person------>loan）

LOAD CSV WITH HEADERS FROM "file:///snapshot/PersonApplyLoan.csv" AS row
FIELDTERMINATOR '|'
MATCH (p1:Person{Id: row.personId})
MATCH (p2:Loan{Id: row.loanId})
CREATE (p1)-[:apply {createTime: row.createTime,org:row.org}]->(p2)

guarantee关系（person----->person）

LOAD CSV WITH HEADERS FROM "file:///snapshot/PersonGuaranteePerson.csv" AS row
FIELDTERMINATOR '|'
MATCH (p1:Person{Id: row.fromId})
MATCH (p2:Person{Id: row.toId})
CREATE (p1)-[:guarantee {createTime: row.createTime,relation:row.relation}]->(p2)

invest关系 (person------>company)

LOAD CSV WITH HEADERS FROM "file:///snapshot/PersonInvestCompany.csv" AS row
FIELDTERMINATOR '|'
MATCH (p1:Person {Id: row.investorId})
MATCH (p2:Company {Id: row.companyId})
CREATE (p1)-[:invest { ratio:tofloat(row.ratio),createTime: row.createTime}]->(p2)

own关系 (person----->account)

LOAD CSV WITH HEADERS FROM "file:///snapshot/PersonOwnAccount.csv" AS row
FIELDTERMINATOR '|'
MATCH (p1:Person {Id: row.personId})
MATCH (p2:Account {Id: row.accountId})
CREATE (p1)-[:own {createTime: row.createTime}]->(p2)

至此，所有关系的导入也结束了。

部分关系的属性我可能没有给他设置数据类型，如果没有特别设置，他默认貌似是string,你们在写的时候，如果注意到某些属性应该设置成什么类型欢迎评论告诉我。