CSV import "updates" unchanged issues

[_mwk]

The customer gave me an XLSX of requirements. I converted it to CSV, used my script to modify the CSV to work as an import to Jira, and then imported that CSV via the Bulk Import option (creation only).

Now the customer has given me an updated XLSX. I converted it to CSV, used my script to modify the CSV to work as an import to Jira (which includes matching their customer's ids with my Jira ids), and then imported that CSV via the External System Import option (to modify existing).

In the updated XLSX, some of the requirements have changed, some have not. But *all* the Updated fields show the time of the External System Import. Looking at the requirements that were the same in both imports, the History shows no change (but as mentioned Updated shows time of import).

Why is Updated being updated when History shows no modification?

[Andy Heinzer · Accepted solution]

Hi

The external system import method is expected to be able to create or update all entries of the CSV file that is being supplied.  So if with the second import you include all the issuekeys or IssueIds of the previous import, and then map to that issuekey/issueId, then Jira will show those issues as having been updated even if no other fields are actually modified.

What is different between the first and second CSV file?  If the latter just has some additional custom field values for some of the issues, but that file still contains all the issuekeys (even those without changes), this could explain the behavior you have seen here.

When mapping to an issuekey, Jira's import still also requires each issue have a Summary field mapped to the import and that field have some value.  Even though those issues might not change the value of the Summary, it is possible if the import file has some different value for that Summary field or any other field mapped to it in the import.  The changelog of the issue in Jira won't show any difference because there is none, but the updated field will be updated to sync with that external system update for any issue that was in the import CSV file.

Does this help explain the behavior seen here?  Curious to learn more about why this might be a problem in this case.

Andy

[_mwk]

I can accept this as an explanation of the behavior.

It is a problem because we use the Updated field to know when to review linked requirements. In requirements management tools, this is sometimes known as ‘suspicion’. If a requirement is linked to another requirement as ‘satisfies’, we’re saying that the first requirement ‘satisfies’ the second requirement, usually with a refinement of detail. If the linked requirement changes, that indicates ‘suspicion’ and we need to review the linking to make sure the change doesn’t affect the other requirement (and make changes if it does). So if nothing changes and the Updated field is updated, then the whole scheme falls apart.

We will probably move to a JSON import using the REST API and use our own diff mechanism.