Progress DataDirect connector for Hadoop Hive from Oracle -

Posted by JIV on 09-Jan-2014 05:36

Hello,

we are evaluating your product - DataDirect for connecting to Hive DB from Oracle DB. So far we managed successfully create DB link from oracle but some types give us strange results.

For example String are interpreted in oracle as Long. I know in Hive there is unlimited size for String but how we could get them as Varchar? Maybe limit size of string in odbc.ini to 4000?

Another thing is DateTime. If such column is selected in Oracle from Hive it shows as random byte stream. Hex preview:  78 6F 07 17 08 01 2F (should be 2011-07-23 07:00:46). And have no idea how get Date from such value.

JV

Posted by Sumit Sarkar on 09-Jan-2014 15:41

Hello JV.  In the odbc.ini, you can set the connection property "MaxVarcharSize" to the maximum length of a string such as "4000".  I have written a blog article on this use case if it's helpful: blogs.datadirect.com/.../instant-oracle-big-data-apps-with-database-gateway-for-odbc-to-hadoop-hive.html

I noticed in our user's guide we support Timestamp data type in Apache Hive 0.8.0 and higher.  

media.datadirect.com/.../wwhelp.htm

What distribution (Cloudera, Apache, HortonWorks, etc) and version are you running?

All Replies

Posted by Sumit Sarkar on 09-Jan-2014 15:41

Hello JV.  In the odbc.ini, you can set the connection property "MaxVarcharSize" to the maximum length of a string such as "4000".  I have written a blog article on this use case if it's helpful: blogs.datadirect.com/.../instant-oracle-big-data-apps-with-database-gateway-for-odbc-to-hadoop-hive.html

I noticed in our user's guide we support Timestamp data type in Apache Hive 0.8.0 and higher.  

media.datadirect.com/.../wwhelp.htm

What distribution (Cloudera, Apache, HortonWorks, etc) and version are you running?

This thread is closed