We have learned in basic database concepts about atomicity, which means doing a complete job or doing nothing. Similarly, if you are exporting data using Sqoop to a table that is very important from the application's point of view and you want make sure that Sqoop should export all data present in HDFS or do nothing, this recipe will help. In this recipe, we are going to see how to ensure atomicity of data export.
To perform this recipe, you should have a Hadoop cluster running with you as well as the latest version of Sqoop installed on it. Here I am using Sqoop 1.4.6. We would also need a MySQL database to be present in the network. Installing Sqoop is easy; by downloading Sqoop tar ball and setting it in the system path. As we are going to import data from MySQL, we would also need to download MySQL connector. Based on your MySQL version, download the right connector jar and copy it into the lib directory of the Sqoop installation.
