Hi justin, I am facing the same problem as qwert as described here in the google group discussion: https://groups.google.com/a/cloudera.org/forum/#!topic/cdh-user/Xc4DUwa2qOQ Unfortunately, even the latest ODBC driver for Impala v2.5.13 does not solve this problem. For background information: I have set up a data source for Hive using the ODBC driver version v2.5.5 (CDH_Hive) and for Impala using v2.5.13 (CDH_Impala). Test connections were both completed successfully. In the SQL Server Management Studio I have created two linked server objects, one for Hive Data source and one for Impala: EXEC master.dbo.sp_addlinkedserver
@server = 'HiveForInitions', @srvproduct='HIVE',
@provider='MSDASQL', @datasrc='CDH_Hive',
@provstr='Provider=MSDASQL.1;Persist Security Info=True;User ID=hive;Password=';
EXEC master.dbo.sp_addlinkedserver
@server = 'ImpalaForInitions', @srvproduct='Impala',
@provider='MSDASQL', @datasrc='CDH_Impala',
@provstr='Provider=MSDASQL.1;Persist Security Info=True;User ID=impala;Password='; Now, I want to query a table named 'yelp_reviews' that was created in Hive Metastore based on the dataset from the yelp dataset challenge. The table only contains int and string columns. The following query for Hive is executed successfully, although it was necessary to set the default string column length to 8000 in the advanced options of the data source (default was 255). SELECT * FROM [HiveForInitions].[HIVE].[default].yelp_reviews Now, changing the query to use the impala data source. SELECT * FROM [ImpalaForInitions].[Impala].[default].yelp_reviews results in the error message: Msg 7399, Level 16, State 1, Line 1
The OLE DB provider 'MSDASQL' for the linked server 'ImpalaForInitions' reported an error. The provider did not give any information about the error
Msg 7306, Level 16, State 2, Line 1
Cannot open the table `Impala`.`default`.`yelp_reviews` from OLE DB provider 'MSDASQL' for the linked server 'ImpalaForInitions'. Unknown provider error. Changing the string column length to 8000 like in the hive data source does not solve the issue either. However, there is no such problem with the impala data source, if the queried table in the metastore only contains int-coulmns. I already verified this as described in the goolge group discussion. Therefore, if I have not overlooked anything, the error results in the handling of string columns. Since this is working for hive data sources, I guess the ODBC driver for impala just needs a little bit of tweaking. Can you please have a look? Thanks a lot and Merry Christmas!
... View more
