Knowledge Base

How to import a GrapheneDB database into Neo4j Desktop

This assumes you already have the zip file of the GrapheneDB database.

The goal here is to import this as an entire database, not to pull the data from it into an existing database.

Note that there are already export instructions on GrapheneDB’s site here: https://docs.graphenedb.com/docs/importing-and-exporting-databases#loading-an-exported-datastore-into-a-local-neo4j-instance

Try using those first, and if those don’t work proceed as follows depending on the Neo4j version of the GrapheneDB database.

In Neo4j 3.5.x

We only have a single database per dbms in this version, so the steps are simpler.

  1. In Neo4j Desktop, create an instance of the right version of 3.5.x in desktop, but don’t start it yet.

  2. Manage the instance (from the …​ in the upper right), and press the Open Terminal button to open a terminal window to the instance’s home directory.

  3. Perform cd data/database from the terminal (if you then do an ls, the directory should be empty).

  4. Extract the GrapheneDB zip file to this directory, make sure that it results in a graph.db folder with neostore* files inside. If needed, rename the top-level directory to graph.db to align with the default graph name (or adjust the graph name in your config file accordingly).

  5. Check ownership permissions for the unzipped directory. They should match with the ownership permissions you see when you do a ls -l at the home directory for the db. If they do not match, you may need to change the ownership of the files recursively so Neo4j Desktop is able to read and write to them: https://linuxize.com/post/linux-chown-command/

  6. Start the database.

In Neo4j 4.x

Neo4j 4.x allows multiple databases per dbms instance. As such we can add the GrapheneDB database without replacing any other databases.

We can even add to an existing dbms instance that already have databases, and that may already be running.

You should have a zipped database, and you may also have a transactions zip too.

  1. In Neo4j Desktop, either create a new dbms instance of the appropriate version, or identify an existing database (whose minor version matches that of the GrapheneDB instance) into which you will import the database.

  2. Manage the instance (from the …​ in the upper right), and press the Open Terminal button to open a terminal window to the instance’s home directory.

  3. Perform cd data/databases from the terminal. This may have other directories here, corresponding with other databases on the instance.

  4. Extract the GrapheneDB zip file to this directory. The top level directory of the expanded zip should have a unique name that follows Neo4j’s naming rules: https://neo4j.com/docs/cypher-manual/current/syntax/naming/

  5. Check ownership permissions for the unzipped directory. They should match with the ownership permissions you see when you do a ls -l at the home directory for the dbms. If they do not match, you may need to change the ownership of the files recursively so Neo4j Desktop is able to read and write to them: https://linuxize.com/post/linux-chown-command/

  6. (Only if you have a transactions zip included) From the terminal, return to the home directory of the Neo4j instance and perform cd data/transactions. This is where transactions for each database are stored. You should see a directory per database.

  7. Extract the transactions zip file to this directory. The top level directory of the expanded zip should have a name matching the database you extracted earlier in step 4.

  8. Check ownership permissions for the unzipped directory. They should match with the ownership permissions you see when you do a ls -l at the home directory for the dbms. If they do not match, you may need to change the ownership of the files recursively so Neo4j Desktop is able to read and write to them: https://linuxize.com/post/linux-chown-command/

  9. Start the Neo4j dbms if it is not started already.

  10. After it starts and you log in, execute :use system to switch to the system database, where we can execute database administration commands.

  11. Using the name of the database (used as the directory name in steps 4 and 7) issue a CREATE DATABASE <dbname> command, replacing <dbname> with the database name. This will create an entry in the system db for the new database, which will be using the database and transaction files extracted from steps 4 and 7.

  12. You can use SHOW DATABASES to check if the database was created.

  13. Execute :use <dbname> (replacing with the database name) to switch to the db.