2 entries are considered duplicate if entry1.db_name == entry2.db_name 
and entry1.db_code == entry2.db_code. The second entry is not added to 
dataEntries vector ( see dataEntries::add() ).

The problem here is that some databases (esp. pir, prf) do not list db
accession code where it should be. Db code is then replaced with "?"
and the following entries for the same database are ignored as they all
have db_code = "?"

Allowing duplicate entries will result in multiple pdb entries when
they list several chains under one accession code (e.g. bmr1000.str).
So we'll have
PDB 1WAV A Chain A...
PDB 1WAV B Chain B...
etc.  

