The phoenix drivers do let *Pentaho to connect to Phoenix via a Generic Database pure JDBC* connection as described in http://talat.uyarer.com/post/121179803796/how-to-connect-hbase-using-apache-phoenix-from
I can see mappings, fields from both sides, source and target! But I can't do an UPSERT into Phoenix from RDBMS as what Pentaho generates on the Target is INSERT SQL statements and the Phoenix driver does not accept that. Pentaho transformation log says *2015/09/22 15:32:37 - Phoenix Table.0 - Caused by: org.apache.phoenix.exception.PhoenixParserException: ERROR 601 (42P00): Syntax error. Encountered "INSERT" at line 1, column 1.* I guess I can write to Pentaho Forums and request *a connector for Phoenix to be build*. They do provide exclusive UPSERT connectors to *Salesforce, MongoDB* etc... This missing Pentaho UPSERT connector for Phoenix (just changing the INSERT to UPSERT) would ease parallel loads into *Phoenix via Pentaho ETL* very much as anybody who uses Pentaho, or even other tools like Informatica would be able to UPSERT directly into Phoenix from RDBMS sources. Best Regards Kubilay Tsil Kara On 22 September 2015 at 11:56, kubilay.tsilkara <[email protected]> wrote: > Hi James > > Thank you! > > I have also found now a way to connect Pentaho PDI (aka Kettle) to > Phoenix. I used steps from this blog > http://talat.uyarer.com/post/121179803796/how-to-connect-hbase-using-apache-phoenix-from > > It works with generic Phoenix drivers, the use case is to create an ETL > transformation from MySQL -> Phoenix to load couple of millions of rows. > The transformation is valid I see the tables on both source and target. > > The hurdle I hit now, is that Pentaho source MySQL Sends INSERTs to > Phoenix, whereas Phoenix expects UPSERTs so I have to do some sort of > transformation in the middle to translate those INSERTs to UPSERTs with a > 'decode' like transformation I suppose. > > Things look very promising so far. > > Cheers. > > Kubilay Tsil Kara > > Kubilay > > On 21 September 2015 at 19:05, James Taylor <[email protected]> > wrote: > >> Have you seen this blog post, as it details how to connect Phoenix >> to Saiku through Pentaho? >> https://blogs.apache.org/phoenix/entry/olap_with_apache_phoenix_and >> >> HTH. Thanks, >> >> James >> >> On Mon, Sep 21, 2015 at 8:39 AM, kubilay.tsilkara < >> [email protected]> wrote: >> >>> I've tried to connect to Phoenix using Pentaho PDI (aka Kettle) >>> <http://community.pentaho.com/projects/data-integration/>with no >>> success. >>> >>> Pentaho is an ETL tool which can do parallel loads to loads of >>> endpoints, including HBase, Hive etc... using JDBC/ODBC connectors. Is >>> there one for Phoenix? Does anybody know? >>> >>> Thank you! >>> >>> Kubilay >>> >> >> >
