You need use sqoop for direct import on hive form sql. First download sql jdbc https://www.microsoft.com/en-us/download/details.aspx?id=11774 place jar on sqoop master server: /usr/hdp/current/sqoop-server/lib use this command for import: import --connect jdbc:sqlserver://[SQL_SERVER_NAME]:[SQL_PORT]/[DB-NAME] --username "[SQL_USERNAME]" --password "[PASSWORD]" --query '[INSERT QUERY HERE] WHERE $CONDITIONS' -m 1 --hive-import --hive-database [DB_HIVE] --create-hive-table --hive-table [HIVE_TABLE] if use --query you must have WHERE $CONDITIONS on query. ... View more

It turned out that the problem was caused by a join with a subquery which made the data to be unevenly distributed among the partitions. Actually I don't know why this is happening but we solved by materializing the subquery. Thank you or the support. ... View more

I found the problem. It was a mistake by me: I added the configuration of the tutorial (https://community.hortonworks.com/articles/4868/hive-oom-caused-by-javalangoutofmemoryerror-java-h.html) via the set command in beeline: instead, they have to be configured in the Tez section in Ambari... ... View more