Update. I have now discovered that this problem is occurring on other node types.
I have done some further diagnosis. This query http://www.screencast.com/t/sYX04qGf4Pc returns a list of User nodes - so far so good. However, when I MATCH for the node by id it finds a different node with the same id - http://www.screencast.com/t/8Ny2QuEVNB So, I delete this folder containing the indexes: http://www.screencast.com/t/32GwWdcara Then watch the indexes repopulate after restarting Neo4j service http://www.screencast.com/t/qQc1RzttHjez When the indexes have repopulated: http://www.screencast.com/t/Zh4CvwvbZ5 I do the search again: http://www.screencast.com/t/gm8ZztMQTK now the MATCH returns two nodes with the same id despite the existence of a Constraint on that property. js On Monday, 3 November 2014 09:22:50 UTC, John Swain wrote: > > I was doing a large number of merge operations adding a new property to > around 5m nodes. > > During testing I noticed that the merge seemed to be creating spurious new > nodes breaking the Constraint condition. It turns out that the index was > out of sync and the merge was not finding the existing record therefore > creating a new one. > > I deleted the Lucene folder and restarted Neo4j - which fixed the problem. > > Here is a screenshot http://www.screencast.com/t/QkfzT1ye > > I would like to know how to avoid this occurring in future and if it is a > known problem. I have searched the group but not found any report if this > issue. This is a pretty serious bug for me as I need to do many of these > merge operations and need to be sure it is reliable. > > I am using Neo4j Enterprise 2.1.5 Personal Edition on a Mac running > Yosemite. Happy to provide any further information if required. > > regards, > js > -- You received this message because you are subscribed to the Google Groups "Neo4j" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
