In Kohadevbox or sandbox, the sample of patrons is not the same if you use MARC21 (53 patrons) or UNIMARC (3 patrons). It could be really great (and easier for tests !) if this sample could be the same, for it's not MARC flavour sensitive datas.
Should be fixed since bug 27619. *** This bug has been marked as a duplicate of bug 27619 ***