In my case, I have to look if my product already exists in a table before insert or update datas, what is the difference between using conditional split or a lookup ? I think the lookup is the best solution because it's just true/false condition (I check the product code to verify if the product is in the table) but my colleague told me the Conditional split is better but not why. Do Conditional split have better performance?
Well, the Conditional Split is useful for working on a dataset and the Lookup is for a simple condition.
For example, imagine I need to work on products from 2016 and another works on products of 2017. In this case I will add a Conditional Split to the Source with conditions Year=2016
and and other line with year=2017
to get 2 flux (2 arrows) containing 2 datasets. The first with 2016 products and the second with 2017 products. In my case, I don't modify the source dataset so I have to use Lookup because it's just a condition.