For more complex operations like updating or removing relationships,
either do not use parallel: true
OR make sure that you batch the work in a way that each subgraph of data is updated in one operation,
such as by transferring the root objects.
If you attempt complex operations, also enable retrying failed operations, e.g. with retries:3.
If this doesn't work, is there anything useful in errorMessages column of apoc.periodic.iterate result and / or in neo4j.log / debug.log files?
@giuseppe_villan I'm on Neo4j Aura so don't have access to neo4j.log unfortunately.
I did manage to get it working by removing the direction on the query. Interesting adding a -> over a - slowed it down. Not sure if you have insight into this.