Head's Up! These forums are read-only. All users and content have migrated. Please join us at community.neo4j.com.
04-20-2020 01:12 PM
I'm getting the following error while trying to load some relationships from the CourtListener dataset:
TypeError: Expected String("https://www.courtlistener.com/api/rest/v3/dockets/12560/") to be a org.neo4j.values.virtual.VirtualNodeValue, but it was a org.neo4j.values.storable.StringWrappingStringValue (Failure when processing file '[REDACTED]/batch.csv' on line 101.)
The import query is:
LOAD CSV WITH HEADERS FROM "file:///batch.csv" AS row
WITH
row.resource_uri AS docket,
row.appeal_from AS appeal_from
WHERE row.appeal_from IS NOT NULL
MATCH (docket:Docket {uri: docket})
MATCH (court:Court {uri: appeal_from})
CREATE (docket)-[rel:APPEAL_FROM]->(court);
and the same resource_uri column from the same CSV was used to create the Docket nodes in the first place, so I'm reasonably confident that it's not that the keys aren't matching up.
The problem seems to be at least somewhat unpredictable. I'm doing the load in batches, and changing the batch size will change what record yields this error, but, for a given batch size, it always has trouble with the record.
Is it possible that it's that I'm loading too many CSVs in rapid succession? I'm converting the data from JSON and loading it in a Python script, using the py2neo library.
attached: screenshot of the EXPLAIN for this query.
All the sessions of the conference are now available online