Use identifier instead of table uid in dataset and transfer table

[brotkrueml]

This is a preparation to #19 to smoothen the transition to version 2.0. The table uid is then not available anymore, instead an "identifier" (formerly known as "handle") is available. The dataset and transfer table should use the identifier/handle instead. This change should be available in version 1.

• A column identifier is added to the dataset and transfer table.
• When processing datasets or transfers the identifier/handle is used instead of the table_uid.
• If an identifier is empty the table uid is used instead.