Luciana,
I've always found it easiest to do the following:
1. Have the HR team prepare a SQL based extract. You'll want them to create the extract based on the minimal dataset needed, as Peter had mentioned, no need to get information that is not needed for the provisioning process. I prefer SQL since the DBAs can make sure that the tables have restricted access to protect privacy. They can also set up secure access to protect the data as well. Something not as easily or comprehensively done with CSV and other text formats.
2. Now you can create an IDM job that will read this data and begin the provisioning process. Simple use this HR table as the source and map it to the relevant IDM attributes in the destination tab.
Note you could also use this for identity updates as well as new user provisioning.
Hope this helps,
Matt
