Generate graph CSVs

Description

The Generate graph CSVs transform (previously Neo4j generate CSVs) writes CSV files with the nodes and relationships of the graphs in a field of the Graph data type: a file per combination of labels and a file per relationship type, in the import/ folder below the base folder.

The headers of the files are for one of these tools:

neo4j-admin import: the headers of Neo4j’s bulk import tool, which the Neo4j import transform runs.
plain column names: just the names, for loading the files into other graph databases, for example with LOAD CSV on Memgraph or with the CSV load functions of Apache AGE.

Options

Option Default Description

Option	Default	Description
Transform name	Generate graph CSVs
Graph field	-	The field in this pipeline that contains the graph data (Graph data type)
Base folder (below `import/` folder)	`/var/lib/neo4j`	the folder to write the nodes and relationships types to. Its `import/` folder is created if needed.
CSV files prefix	prefix	prefix to apply to the generated CSV files
Node/Relationships Uniqueness strategy	None	* None: don’t handle uniqueness while generating CSVs * First: if duplicates are found, keep the first occurrence * Last: if duplicates are found, keep the last occurrence Please note that primary key of all nodes written to a CSV file needs to be unique. The uniqueness strategy used in this transform is simple in nature and loads everything in memory. Because of this you can consider using a Sort transform to force uniqueness beforehand.
CSV header format	neo4j-admin import	* neo4j-admin import: `id:ID,name:string,…,:LABEL` for nodes and `:START_ID,…,:END_ID,:TYPE` for relationships. * Plain column names: `id,name,…,label` for nodes and `start_id,…,end_id,type` for relationships. A `Vector` property is written with `;` between its numbers, as a `float[]` column in the neo4j-admin import header. The rows are the same with both.
Filename field	filename	field that contains the filename for the CSV files to be generated
File type field	fileType	field that contains the file type for the CSV files to be generated

Transform name

Generate graph CSVs

Graph field

The field in this pipeline that contains the graph data (Graph data type)

Base folder (below import/ folder)

/var/lib/neo4j

the folder to write the nodes and relationships types to. Its import/ folder is created if needed.

CSV files prefix

prefix

prefix to apply to the generated CSV files

Node/Relationships Uniqueness strategy

None

* None: don’t handle uniqueness while generating CSVs * First: if duplicates are found, keep the first occurrence * Last: if duplicates are found, keep the last occurrence

Please note that primary key of all nodes written to a CSV file needs to be unique. The uniqueness strategy used in this transform is simple in nature and loads everything in memory. Because of this you can consider using a Sort transform to force uniqueness beforehand.

CSV header format

neo4j-admin import

* neo4j-admin import: id:ID,name:string,…,:LABEL for nodes and :START_ID,…,:END_ID,:TYPE for relationships. * Plain column names: id,name,…,label for nodes and start_id,…,end_id,type for relationships.

A Vector property is written with ; between its numbers, as a float[] column in the neo4j-admin import header.

The rows are the same with both.

Filename field

filename

field that contains the filename for the CSV files to be generated

File type field

fileType

field that contains the file type for the CSV files to be generated