in connection with my work on Pentaho (5.0.1.) and my Hadoop Cluster I am confronted with several problems. The Hadoop Cluster is based on the newest Cloudera all-in-one VM (CDH 4.4, Impala 1.1.1), but the use of the database connection via the Impala connector causes some difficulties with regard to different data types.

Using the database connection via Impala connector, I got some problems regarding different data types. If I try to read in (table input via Impala connector) some double values stored in a Impala created table, they always get rounded to just one position after decimal point. No error message. All other data types work fine.

Furthermore, when I try to write some double values via a table output and via the Impala connector, Pentaho is reporting an error message such as “Method not supported”. If required, I can send you the entire/complete error log.

What is more, there are some difficulties concerning the writing of timestamp data types to Impala (again table output via Impala connector). However, this issue seems to be a more obvious one. When I try to write via the “insert to” command, the >’< sign around the timestamp value is missing. The way it would work, would be the following: insert into myTable (myDate) values ('2014-01-08 13:43:12.0') Writing String and Int works fine.

Is there any possibility to get this work?
Many thanks in advance for your help.