We are using OrientDB as document database in plocal mode. Classes are schema-mixed. We manage cluster selection by some record fields such as timestamp, using doc.save(clusterName).
When looking inside cluster PCL files I find that every record has the class name as a string. This has a massive impact on the size of the file when the record is small.
Why is class name saved with every record? The cluster belongs to the class by oClass.addCluster(clusterName), so maybe the class name in the record is redundant?
Thanks, Chris