incorrect mapping between bitstreamforamt and fileextension

[Paurikova2]

We have two tables:

• bitstreamformatregistry: id, mimetype, short_description, ...
• fileextension: id, bitstream_format_id, extension.

The mapping between these tables is by bitstreamformatregistry id. The attribute short_description is unique.

Problem: In DSpace 7, new records were added to the bitstreamformatregistry table. We imported bitstreamformatregistry but without fileextensions. If data already exists in the database (short_description is not unique), the database throws an exception and we log the info. However, there are records where short_description values differ only in lower/upper case.

Problems:
Records in bitstreamformatregistry are imported without fileextensions.

Question:

• Should we have a unique mimetype?
• What if one mimetype in DSpace 5 had different extensions than in DSpace 7?

Test:
We need to validate the mapping, not only the id, but all columns must be correct because of the implicitly imported data!