I want to migrate from oracle database to hadoop using python scrip without Sqoop & Spark .What is the easy way for migrating .
You could connect to the oracle database from python using any interface. Lets say - cx_Oracle from - 'https://www.oracle.com/technetwork/articles/dsl/python-091105.html'. You will now be left with crafting the python script to run in a multi process fashion on a single machine to get as much data as possible quickly. The next step will be to make use of all the available machines in your cluster to do that job for you.
The tool - sqoop cropped out of the requirements of dumping loads of data between Hadoop to other RDBMSs. The tool handles creation of multiple map/reduce jobs to achieve the right performance and resource utilization.
Sqoop could also create tables in Hive so as to save you those steps.
Any reason you would be unable to use sqoop to save most parts of the exercise?