Many times I find conference presentations when included in journals all have the same DOI. The abstracts are listed separately and are provided to databases as different items. However, the items do not have different DOIs. This is unfortunate.
Ideally, DOIs are unique identifiers. However, (especially with ahead-of-print articles), not always. Here are 4 of a recent 20 examples. There is little that Zotero can do about this. (Okay, the F1000 example may not be the best.)
That's a different case though: As per their specifications, DOIs must uniquely identify an object, i.e. the same DOI cannot point to multiple objects. DOIs also should be unique in identifying an object, i.e. there should only be one DOI referring to the exact same object, but this is not strictly required.
10.1177/00031348231161756
10.1177/00031348231175099
10.1556/650.2023.32819
10.1556/650.2023.32786
10.1542/peds.2023-063255
10.1542/peds.2023-063256
10.12688/f1000research.111002.1
10.12688/f1000research.111002.2
Translated version of article vs. vernacular
10.1016/j.ad.2022.12.007
10.1016/j.ad.2023.10.018
English edition of journal vs. vernacular edition
10.1016/j.nrl.2020.03.003
10.1016/j.nrleng.2020.03.005
Same article by institutional author in different journals
10.1016/j.nwh.2022.09.001
10.1016/j.jogn.2022.09.002
As per their specifications, DOIs must uniquely identify an object, i.e. the same DOI cannot point to multiple objects.
DOIs also should be unique in identifying an object, i.e. there should only be one DOI referring to the exact same object, but this is not strictly required.
Publishers don't always behave as expected.