Duplicate records found in kb_a_addr_inst.csv

[saschaszott]

We (University Library of Helmut-Schmidt-University Hamburg) extracted the records from the latest kb_a_addr_inst.csv (taken from https://zenodo.org/records/18429476/files/kb_a_addr_inst.csv?download=1) that are linked to our institution (ID 136).

Out of 8,999 records, we found duplicates. Only 8,581 records are unique based on the OpenAlex Work ID.

For example, these two records in the file appear as duplicates:

136,https://openalex.org/W997221190,"Helmut-Schmidt-Universität/Universität der Bundeswehr Hamburg, Hamburg, Deutschland",uni,10.1007/978-3-658-02620-2_12,10.1007/978-3-658-02620-2_12§136
136,https://openalex.org/W997221190,"Helmut Schmidt Universität/Universität der Bundeswehr Hamburg, Hamburg, Germany",uni,10.1007/978-3-658-02620-2_12,10.1007/978-3-658-02620-2_12§136

In another case, we even found 4 records with the same OpenAlex Works URL:

136,https://openalex.org/W4382050566,"Helmut Schmidt University Hamburg,Institute of Automation Technology,Hamburg,22043",uni,10.1109/icuas57906.2023.10155997,10.1109/icuas57906.2023.10155997§136
136,https://openalex.org/W4382050566,"Helmut Schmidt University Hamburg,Institute of Public Law,Hamburg,22043",uni,10.1109/icuas57906.2023.10155997,10.1109/icuas57906.2023.10155997§136
136,https://openalex.org/W4382050566,"Helmut Schmidt University Hamburg,the Institute of Automation Technology,Hamburg,22043",uni,10.1109/icuas57906.2023.10155997,10.1109/icuas57906.2023.10155997§136
136,https://openalex.org/W4382050566,"Helmut Schmidt University Hamburg,the Institute of Public Law,Hamburg,22043",uni,10.1109/icuas57906.2023.10155997,10.1109/icuas57906.2023.10155997§136

Could you clarify why these duplicates occur? Are they expected due to multiple affiliation records or different metadata sources in OpenAlex?

[naustica]

Hi @saschaszott ,

this is probably because the (OpenAlex) address strings in the examples differ slightly and the algorithm can only assign unique addresses. Once a comma, semicolon, etc. is included, it is considered a new address for the same work id. I hope this helps.

Best,

Nick