How can I import data into Neo4j?

Approaches for Importing Data

When I say importing data, I mean the process of moving data from one system into Neo4j.

The import could be a one-time operation, or it could be a regular process that you need to automate.

Typically, the source system would have a different data model than Neo4j, and by importing data, you would transform the data into a graph model.

The source may expose data in different ways, for example:

Tools

There are many tools available to import data into Neo4j, including:

Cypher and LOAD CSV

Cypher has built-in support for importing data from CSV files using the LOAD CSV clause.

LOAD CSV WITH HEADERS FROM 'file:///transactions.csv' AS row

MERGE (t:Transactions {id: row.id})
SET
    t.reference = row.reference,
    t.amount = toInteger(row.amount),
    t.timestamp = datetime(row.timestamp)

You can control the import process by writing Cypher queries to:

neo4j-admin

The neo4j-admin import command line interface supports importing large data sets. neo4j-admin import converts CSV files into the internal binary format of Neo4j and can import millions of rows within minutes.

The neo4j-admin import command expects you to format the data in a specific way and requires the database to be offline during the import process.

Data should preferably be clean and transformed before importing, as faults in the data will reduce the performance of the import process.

The neo4j-admin import is a highly configurable high-performance tool, applicable when you need to import large data sets very quickly.

ETL (Extract, Transform, Load) Tool

An ETL tool, for example Apache Hop, is a good choice for importing data from multiple sources. ETL tools generally support various data sources, can transform data into the desired format, and have visualization tools.

Many organizations use ETL tools to import data into Neo4j because they can handle complex data transformations and integrations.

Custom integration using Neo4j drivers

Building a custom application to load data into the graph database is a good option if you have complex business rules or need to integrate with other systems. A custom application will allow you complete control over the import process and integration with other systems and data sources.

There are several GraphAcademy courses for developers, where you can learn how to build applications using Neo4j drivers.

Mixed approaches

In practice, you may use a combination of these tools to import data into Neo4j. For example, you may use Data Importer for quick prototyping, Cypher for small data sets, and ETL tools for complex data transformations.

You may also choose to do one-off batch imports using one tool and real-time data ingestion using another tool.