An updated picture of the ligand sets and copper-ligand atom bond lengths in proteins is presented which takes advantage of (i) the approximately twofold increase in the number of entries for copper-containing proteins in the PDB since the last study of this kind, especially benefiting from the recent incorporation of the structures of proteins involved in copper homeostasis, and (ii) a preliminary classification of copper sites based on their structural, electronic and functional features. This classification allowed the calculation of reliable target copper-ligand distances for several bonds that were not available in previous work and that are in good agreement with EXAFS data and the known chemistry of these sites. The analysis presented here further disclosed an artifactual dependence of the average of the reported Cu-NHis bond lengths on structure resolution, highlighting the importance of taking this into account when computing target distances even from high-resolution structures. Finally, a relationship between the two Cu-O distances in bidentate carboxylates is disclosed, similar to that reported previously for other metal ions.