xghrbc1001/neo4j-rdbms-import

Massively parallel RDBMS to Neo4j Importer

Relational to Neo4j Importer

Scans Database Metadata to determine how to model Uses Parallel Batch Importer to ingest data from all tables in parallel

Blog Post: docs/relational_to_neo4j_import_tool.adoc

Usage

Currently you need apache maven installed.

Command line parameters are:

import-rdbms-mvn.sh "jdbc:url" "schema" "graph.db"

e.g.

import-rdbms-mvn.sh "jdbc:mysql://localhost:3306/northwind?user=root" "northwind" "northwind.db"

Transformation Rules

• Table with primary key become nodes

• FK (1:1, 1:n and n:1) become relationships

• Tables with no PK and exactly two FK and properties become relationships (optionally also if there is a pk)

• Tables with more than two FK become nodes

• remove all fk fields

Todo

• Table name, column name transformation

• Modeling Rules by table name pattern

• Type transformation

• Skipping Columns / Rows

• Unique Constraints

• Indexes

• Compound Indexes (create artificial property)

• Skip keys

• Limit row rounds, by time

• Custom SQL statements

• Configure concurrency for JDBC reads

Ideas

neo4j-rdbms-import --into …​ --relationship-rule no-pk, two-fk, table% | default: no-pk --skip table% table%.field,table_.id | default none --rename: label: Capitalize,UP_CASE,camelCase,CamelCase | default:CapitalizeCamel relationship: property:

• include and exclude with patterns (% and _)