Want to use kettle to achieve a very simple requirement, from MySQL to Oracle import a table data, if the table in Oracle does not exist, first table and then import data. This is a seemingly simple feature, but it may be a bit confusing for users who have just started to touch kettle. There is a "check table exists" step in kettle conversions and assignments, but if you want to achieve the above requirements, you can only use the job, not the conversion.
In kettle, the job is serially executed, and the next job item starts only if the previous job item succeeds. Multiple steps in a transformation are executed in parallel, and the data flow operates concurrently in all the steps that make up the transformation. Therefore, if you use transformations, you cannot guarantee the order in which the tables are to be built before the data is imported. Here is the experimental procedure.
1. Create a job called "MySQL to Oracle", as shown in the following illustration.
2. The "Check table exists" Job item is shown in the following figure.
3. The "CREATE TABLE" SQL Script job item is shown in the following figure.
4. The "MySQL to Oracle" conversion is shown in the following three images.
Reference:http://www.it4biz.com.br/cursos/kettle-cookbook/Samples/html/jobs/Slowly%20Changing%20Dimension/ Create%20-%20populate%20-%20update%20slowly%20changing%20dimension.kjb.html