Neo4j > Basics > Data Import and Export

---

In case you have a huge data set that you want to insert into Neo4j at once such as when you want to migrate from another datasource, Cypher language supports loading data as parameters. These parameters can be either scalar values, maps, lists or lists of maps. The UNWIND clause can be used to help expanding the imported data and inserting it one by one to create the graph structure. You can also import easily a JSON data pulled from some API as seen below:

The products JSON data:

{
  "products" : [ {
    "name" : "T-Shirt",
    "size" : 20,
    "price": 200 
    },
    {
    "name" : "Laptop",
    "size" : 15,
    "price": 1200 
  } ]
}

Then imported as:

UNWIND {products} as product
MERGE (p:Product {name:product.name}) ON CREATE SET p.created = timestamp()

Another option to import huge set of data is by importing the data from csv files. Cypher supports the LOAD CSV clause that will parse a local or remote file into sequence of rows, then you can use Cypher clauses and operations to create the desired nodes and relationships. Usually it would be better if you load the nodes and relationships from separate CSV files. The example below shows how to load a CSV file contains a set of products with the format below:

id,name,size,price
1,T-Shirt,20,200
2,Laptop,15,1200
3,Iphone,5,600
4,Samsung 5s,4,400

Load the data as shown below:

LOAD CSV WITH HEADERS FROM "http://{url}/products.csv" AS line
MERGE (p:Product { id:line.id })
ON CREATE SET p.name=line.name,p.size=line.size,p.price=line.price;

Then to create relationships, we can load the orders and create the relationship between each order and product node during loading as seen below:

First load the orders:

LOAD CSV WITH HEADERS FROM "http://{url}/orders.csv" AS line
MERGE (o:Order { id:line.id })
ON CREATE SET o.status=line.status,o.total=line.total;

Then load the OrdersProducts table

LOAD CSV WITH HEADERS FROM "http://{url}/OrdersProducts.csv" AS line
MATCH (p:Product { id:line.productId })
MATCH (o:Order { id:line.orderId })
CREATE (o)-[:Contains]->(p);