Thanks! About the Long vs long, indeed after experimenting with changing all Long types to long in the generated code I found out that there was a problem when it tried to set NULL on a primitive long. So they should remain Long where NULL is allowed as you indicate indeed (otherwise it would cause problems with the SQL queries if it would use for example - since that would invalidate the use of the IS NULL clause).
The reason why I got on that long track is that I get an error (after correcting the index uppercase and date/timestamp in code) in 'BinaryUtils.mergeMetadata' when it tries to merge two values for the 'nameMlid' field. At a certain point it generates a mismatch error because newFieldTypeName is set to long and OldFieldTypeName is set to Object. Since the field is typed as 'Long' and it is displaying 'long' there I assumed that the problem orignated from the types used on the definition - which now doesn't seem to be the case. So I will have to add another question I'm afraid, any idea why it is being confused here. Attached a log with the full trace as well. binaryMismatch.log <http://apache-ignite-users.70518.x6.nabble.com/file/n3813/binaryMismatch.log> I'll have a go with the web console, if I can get it working through our proxy that is. In the end I will probably be generating meta data using our own meta repository but this looks like a nice feature to kickstart indeed. Thanks again -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Ignite-Schema-Import-Utility-Mismatch-when-loading-data-between-Date-and-Timestamp-tp3790p3813.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
