Hi, sunerhan

As I know, Carbon support upward compatible,that means if we create carbon
table in spark1.6 + carbon1.1.0,we can get access this created table in
spark2.1 + carbon1.1.0.

Actually, Carbon will do schema path check in spark 2.1 + carbon1.1.0.
Besides, Carbon use different CarbonDatasourceHadoopRelation for integrating
spark1.6 and spark2.1. This part of process logic has been refactored.

Carbon can't identify your carbon table created in spark2.1 + carbon1.1.0
while you query it in spark1.6 + carbon1.1.0, it may be treated as hive
table, and if the path is not exists or the schema file is incorrect, then
error will happen.

Regards.
Chenerlu





--
View this message in context: 
http://apache-carbondata-dev-mailing-list-archive.1130556.n5.nabble.com/SPARK-update-error-tp13144p13211.html
Sent from the Apache CarbonData Dev Mailing List archive mailing list archive 
at Nabble.com.

Reply via email to