CODEWITHSUNDEEP
csvimportneo4jcypher
Rajat
Rajat

Reputation: 107

Update Data in Neo4j

I am querying data from MSsql server and saving it into CSVs. With those CSVs, I am modelling data into Neo4j. But Mssql database updates regularly. So,also wants to update Neo4j data on the regular basis. Neo4j has two types of nodes: 1.X and 2.Y . Below query and indexing, used to model the data:

CREATE INDEX ON :X(X_Number, X_Description,X_Type)
CREATE INDEX ON :Y(Y_Number, Y_Name)

using periodic commit
LOAD CSV WITH HEADERS FROM "file:///CR_n_Feature_new.csv" AS line           
Merge(x:X{ 
X_Number : line.X_num,
X_Description: line.X_txt,
X_Type : line.X_Type,
})
Merge(y:Y{ 
Y_Number : line.Y_number,
Y_Name: line.Y_name,
})
Merge (y)-[:delivered_by]->(x)

Now there are two kinds of updates:

  1. There may be new X and Y nodes which can be taken care by "Merge" command.
  2. But there can be modifications in the properties of already created nodes X and Y....for e.g. if node X{X_Number : 1, X_Description : "abc", X_Type : "z"} but now in updated data X node properties got changed to X{X_Number : 1, X_Description : "def", X_Type : "y"}

So I don't want to create a new node for the X_Number:1 node but just want to update the existing node properties like X_Description and X_Type.

Upvotes: 2

Views: 620

Answers (1)

Dave Bennett
Dave Bennett

Reputation: 11216

You could just re-write your query to support new nodes and changes to existing nodes by merging only on the X_Number or Y_Number attributes.

USING PERIODIC COMMIT
LOAD CSV WITH HEADERS FROM "file:///CR_n_Feature_new.csv" AS line           
MERGE (x:X {X_Number: line.X_num})
SET X_Description = line.X_txt, X_Type = line.X_Type
MERGE (y:Y {Y_Number: line.Y_number})
SET Y_Name=line.Y_name
MERGE (y)-[:delivered_by]->(x)

This way the MERGE statements will always match the existing X and Y nodes based on the X_Number and Y_Number attributes which presumably are immutable. Then the existing Description, X_Type and Y_Name attributes will be updated with the newer values.

Upvotes: 1

Related Questions