Is it better to use one data transform or many?

Question

We're essentially bulk loading data from one database to another, using SSIS. My co-worker has made one data transform object, and within it, he has a transform for each table. I've made a transform object for each table. We're both rather new to SSIS. Is one way better than the other? What are advantages or disadvantages to each method?

Answer 1

With one data transform object can you readily (re)load just one table?

Data migration exercises don't always work first time. If you need to experiment, adapt and retry a portion then finer granularity of control may be very helpful. Down side of fine granularity is that you have more pieces to manage.

Answer 2

If you meant, multiple data flows inside a Data flow task (case A) and multiple data flow tasks to handle different table (case B), the performance will be somewhat the same, if you allow them to execute in parallel. But if you want to consider other things such as error handling, package restartability and all I recommend case B. Because SSIS does not support restartability inside data flow task. And it will be better if you take a look at the SSIS architecture, synchronous and blocking transformations, etc before you decide.