Neo4j Versioner SQL Documentation

Neo4j Versioner SQL is a collection of procedures, aimed to help developers to manage a SQL Database schema, through the Entity-State model applied to Neo4j.

License

Apache License 2.0

Installation

1. Neo4j Versioner Core is required, please see the current documentation here;
2. Download the latest release;
3. Put the downloaded jar file into $NEO4J_HOME/plugins folder;
4. Put your database JDBC driver into $NEO4J_HOME/plugins folder, for more information on supported databases, click here;
5. Start/Restart Neo4j.

About

Neo4j Versioner SQL has been developed by Alberto D’Este and Marco Falcier.

Data Model

The current data model is a hierarchical representation of the database, that recursively uses the Core entity-state model. Entity nodes are:

• Database: it has a type (postgres, mysql, …) and a name properties;
• Schema: it has a name property;
• Table: it has a name property.

Every Entity has its own State nodes: in Table’s State node we can find all columns name as keys, and their properties as values, for example

id: int
name: varchar

Table’s State nodes can be connected together by the RELATION relationship, which mean that one of the two connected Table’s State node, has at least one foreign key to another Table’s State node. Every State node, with its RELATION relationship is a snapshot of the current structure of the SQL database.

There are 3 different relationships, in addition to the Core ones:

• (:Database)--(:State)-[:HAS_SCHEMA]->(:Schema)
• (:Schema)--(:State)-[:HAS_TABLE]->(:Table)
• (:Table)--(:State)-[:RELATION {destination_column: id, source_column: external_id, constraint: external_id}]->(:State)--(:Table): it contains the foreign key information.

This is how the data model looks like:

Currently supported databases

• postgres
• mysql

Procedures Reference

Neo4j procedure documentation can also be found using CALL dbms.procedures().

Procedure CheatSheet

Here you can find a “compressed” list of all the procedures:

Legend

• Optional parameter
• Node/Path

init

This procedure is used in order to initialize a Database Version with its schemas, tables and relative columns. It will create a graph as showed in the data model section.

Details

Name

sql.versioner.init

Parameters

Return value

Example call

CALL sql.versioner.init('mysql', 'localhost', 3306, 'new_schema', 'root', 'password') YIELD node return node

reload

This procedure is used in order to reload a new Database Version with its schemas, tables and relative columns. It will create a new State for every entity change, if needed.

Details

Name

graph.versioner.init

Parameters

Return value

Example call

CALL sql.versioner.reload('localhost', 3306, 'root', 'password') YIELD node return node

Feedback

We would appreciate your feedback about our Versioner SQL, how to improve and fix (we hope not so many!) any bad things. Say yours in the issue section.