Sqoop vs Oracle Hadoop Connectors

Question

I have used Sqoop to ingest data from Oracle to Hadoop and it worked well but I've no idea how it will work in case of Oracle Hadoop Connector. It took only 3 mins to bring 76 million records from Oracle to Hive table without using partitions on Sqoop.

Can anyone give some details about Oracle Hadoop connectors, Will it perform better than Sqoop?

Answer 1

Most of the connectors would have the performance close to same as you'll have a set of MapReduce jobs on the very end of your workflow and this would play the main role in your overall performance.

Oracle provides a set of different connectors for accessing the Hive and you could check a nice overview about standard solutions 
but I doubt that on the very end you will expect significant performance differences other then you see in Sqoop:

https://docs.oracle.com/cd/E37231_01/doc.20/e36961/start.htm#BDCUG119

Personally, I prefer Sqoop for Hadoop-driven import-export operations and connector approach for querying the data in Hadoop.

Hope this will answer your query to some extent.