Aside from support for CDH4, are there any new features with the new connectors?
I reviewed the data sheet:
But it doesn't differentiate old and new features very well...
It sounds like the direct connector (AKA SQL connector) now automatically creates the external table? Is that right?
What else is new?
Yes, the Oracle SQL Connector for HDFS (renamed from the Oracle Direct Connector for HDFS) can now automatically create the external table. It can also automatically manage the location files - for example, if there are new data files the user can update the locationFileCount parameter and automatically modify the external table to create new location files.
The other key feature is the ability to access data directly from Hive tables. Oracle SQL Connector for HDFS now enables Oracle SQL queries on Hive table data via external tables. The external table is automatically generated based on Hive metadata.
Other new features include performance enhancements with the support of SDP (sockets direct protocol) over Infiniband and support for CDH4. Oracle Loader for Hadoop supports new input formats (the regular expression input format will help in the automatic parsing of a variety of weblog formats).