Closed kd-ods closed 2 weeks ago
Methodology: Counted all statements by recordId_recordType_recordStatus and flagging those that have more than one instance where recordStatus = new Might not find everything if recordStatus incorrectly set.
bods-package-entity-owning-entity.json:
recordId: 12b7dd0770ce ("JENEX LIMITED", "MVJ LIMITED")
bods-package-fi-soe.json:
recordId: 19f1c5afe9d7 ("Gasgrid Finland Oy", "Suomen tasavalta")
joint-ownership.json:
recordId: 31c55e425764 ("CHRINON LTD", "Joint shareholding")
bods-package-annotations.json:
recordId: 387a14452645 ("MONSOON HOLDINGS (NO.1) LIMITED", "DRILLGREAT LIMITED")
No sign of any relationships that have this problem though.
Noting here that incorrect recordIds will skew the 'structure' of the networks. Here is how the networks should look (i.e what these examples looked like in their BODS 0.3 form):
bods-package-entity-owning-entity.json
bods-package-fi-soe.json
joint-ownership.json
bods-package-annotations.json
Note: I made basic network visualisations of all the BODS example data to check see if any edges were broken due to recordId problems. And apart from the files notes above, all looks fine.
Summary of the bug or issue
Some of the files in the examples folder (on main, but also for the 0.4 release) are incorrectly re-using recordId values for different entities. All the example files need to be checked and fixed.
Suggested resolution
@radix0000 to identify which files need fixes.
@kd-ods - fix these files on the main branch.
(I suggest that we hold off fixing them on the 0.4 release branch: I suspect that we will uncover other problems with the example data once the new checks are in the Data Review Tool. Eventually we can fix up all the 0.4 example data on main, then apply the fixes to the 0.4 release branch too.)