CSV LOAD and updating existing nodes / creating new ones

2
votes

I might be on the wrong track so I could use some helpful input. I receive data from other systems by CSV files which I can import into my DB with CSV LOAD. So far so good.

I stucked when I need to reload the CSV again to follow up updates. I cannot delet the former data as I might have additional user input already attached so I would need a query that imports the CSV data, makes a match and when it finds the node it will just use SET to override the existing properties. Saying that I am unsure how to catch the cases where there is no node in the DB (new record) and we need to create a node.

LOAD CSV FROM "file:xxx.csv" AS csvLine
MATCH (c:Customer {code:"ABC"})
SET c.name = name: csvLine[0]
***OPTIONAL MATCH // Here I am unsure how to express when the node is not found***
MERGE (c:Customer { name: csvLine[0], code: csvLine[1]})

So ideally Cypher would check if the node is there and make an UPDATE by SET the new property coming with the CSV or - if the node cannot be found - creates a new one with the CSV data.

And - as a sidenote: How would I find nodes that are not in the CSV file but in the DB in order to mark them as obsolete? (This might not be able in the import but maybe someone has an idea how to solve this in order to keep the DB clean of deleted records - which can only be detected by a comparison with the latest CSV import - happy for every idea).

Any idea or hint how to write the query for updaten the graph while importing?

1
csvneo4jcypher

1 Answers

3
votes

You need to use MERGEs ON MATCH and/or ON CREATE handlers, see http://neo4j.com/docs/stable/query-merge.html#_use_on_create_and_on_match. I assume the customer code in the second column is the identifier - so the name in column one might change on updates:

LOAD CSV FROM "file:xxx.csv" AS csvLine
MERGE (c:Customer {code:csvLine[1]})
ON CREATE SET c.name = csvLine[0]  
ON MATCH SET c.name = csvLine[0]