search

pentaho / mondrian

Mondrian is an Online Analytical Processing (OLAP) server that enables business users to analyze large quantities of data in real-time.
http://mondrian.pentaho.com/
Other
1.14k stars 722 forks source link

How to integrate mondrian with hive-thrift server? #764

Open Zhangshunyu opened 8 years ago

Zhangshunyu commented 8 years ago

Hi, I have a question that how can I integrate mondrian with spark-sql using hive-thriftserver? Thanks for any reply.

mkambol commented 8 years ago

The Apache hive jdbc driver can be used to connect to SparkSQL. There's also a Simba driver specifically for Spark: http://www.simba.com/drivers/spark-jdbc-odbc/

Note that there's not a Mondrian dialect specifically for SparkSQL, so it's possible that you'll hit some issues.

Zhangshunyu commented 8 years ago

@mkambol thanks for your kindly reply, now I intrgrated spark with hive jdbc, and I want to intrgrate them with mondrian, I mean we use Mondrian to send the sql to JDBC, and using hive to connect with spark, just like beeline, is it possible?

mkambol commented 8 years ago

Mondrian supports jdbc connections, so you should be able to define a connection much like you do with beeline to connect. As I mentioned there's been little testing of Mondrian with SparkSQL, and there's no defined dialect, so you may run into issues, but it should be possible to connect.

WangTaoTheTonic commented 8 years ago

Hi @mkambol , I've connected SparkSQL Thrift Server using hive jdbc driver. We wanna change configurations in hive-site.xml which is under under "$TOMCAT_HOME/webapps/mondrian/WEB-INF/classes" but found the jar containing driver class could not read the hive-site.xml.

Do you guys have any ideas about that? Is there sth wrong about what we've done? Note: the log4j.properties under the same directory can work.

WangTaoTheTonic commented 8 years ago

We've solved that. It is right to put it under that directory. The reason is our mistake on the other perspect.