Tag: graph databases

I’ve been working to better understand graph databases and where they can be useful.

There is a file from Neo4J that comes with the Desktop and contains a data export from Northwind. This looks like this when you open it:

On the page where Neo4j talks about the employee imports, they have this code:

// Create employees

LOAD CSV WITH HEADERS FROM 'https://gist.githubusercontent.com/jexp/054bc6baf36604061bf407aa8cd08608/raw/8bdd36dfc88381995e6823ff3f419b5a0cb8ac4f/employees.csv' AS row

MERGE (e:Employee {employeeID:row.EmployeeID})
   ON CREATE SET e.firstName = row.FirstName, e.lastName = row.LastName, e.title = row.Title;

This loads fine.

When I run this code, it works. When change to this, it fails. It’s erroring out on my file. I tried moving the file, removing nulls, etc.

LOAD CSV WITH HEADERS FROM 'file:///employees.csv' AS row
MERGE (e:Employee {employeeID:row.EmployeeID})
ON CREATE SET e.firstName = row.FirstName, e.lastName = row.LastName, e.title = row.Title;

Here’s the error in the Neo4j Browser:

I kept editing the file and trying different things. I compared what I had locally with what was on GitHub. Eventually, I realized this is the issue:

{employeeID:row.EmployeeID}

In the GitHub csv, the first row has headers with EmployeeID. In my local file, the header is “employeeID” (lower case). As soon as I edited this, it worked.

THIS IS THE FILE THAT CAME WITH NEO4J DESKTOP.

They make mistakes, just like I do. I hate that there are too many things like that that are care sensitive. I get that sometimes you want data to be in a particular format, but I think in the modern world, making metadata case sensitive isn’t productive.

I was doing a little experimenting with graph databases. My goal was to run Neo4j in a docker container, but it wasn’t quite as simple as I expected. I started with a How-To, but it wasn’t enough. This post covers a few things I did to get this working.

In looking at the How-To, I got this command:

docker run \
     --name testneo4j \
     -p7474:7474 -p7687:7687 \
     -d \
     -v $HOME/neo4j/data:/data \
     -v $HOME/neo4j/logs:/logs \
     -v $HOME/neo4j/import:/var/lib/neo4j/import \
     -v $HOME/neo4j/plugins:/plugins \
     --env NEO4J_AUTH=neo4j/test \
     neo4j:latest

That’s the basic command I was looking for to run. I tried it, but this doesn’t paste well into a Windows shell. I put the entire thing in a .CMD file to run, making it one long line in the file.

However, when I ran it, I got this:

What’s wrong? Well, $Home is a Linux concept, and it’s looking to map some volumes to folders. The Windows equivalent is %HOMEPATH%. If I check that variable, I see it’s /Users/Steve on my machine.

I created the folders needed for the four volumes, as you can see here:

I tried again, but this didn’t work. I realized, I needed to edit $HOME to %HOMEPATH%. I should have done that immediately.

The next attempt has me seeing this:

Of course that isn’t the right path. I added a “C:” to each volume and things worked.

The final command (For me):

docker run --name testneo4j -p7474:7474 -p7687:7687 -d -v c:%HOMEPATH%\neo4j\data:/data -v c:%HOMEPATH%\neo4j\logs:/logs -v c:%HOMEPATH%\neo4j\import:/var/lib/neo4j/import -v c:%HOMEPATH%\neo4j\plugins:/plugins --env NEO4J_AUTH=neo4j/test     neo4j:latest

If you want to play with Neo4j and don’t want to install things, you can try this. Of course, you still need a query method from an app or a client tool.