Valid Issue key not required to update an Issue through CSV import

Summary

When importing a CSV file and the Issue key is defined as a column, in the Issue key does not exist, JIRA will "assume" the Issue and update accordingly based on the Issue number.

Steps to Reproduce

1. Have an existing Issue in JIRA (ie: SCRUM-379)
2. Prepare a CSV file to update any given field (ie: Summary) using an Issue key that does not exist

   IssueKey, Summary
   RUBBISH-379, Summary update 1
   

3. Import the CSV file and make sure the Issue key is mapped
   

Please note, this bug can also impact if the issue key is moved under a different project previously. ie. ABC-100 is moved as XYZ-200. The CSV import having issue number for UVW-100 will eventually update XYZ-200 to ABC-200.

Expected Results

Issue key does not exist, I would expect the import to fail or do nothing.

Actual Results

JIRA ignores the Project key (SCRUM) and looks for whatever Issue number it can match with (379) and updates that Issue.

Notes

In the example I provided, I imported a CSV file to update the Summary field of Issue RUBBISH-379. This Issue does not exist and there is no RUBBISH Project on this instance.

The importer simply looked at the Project it was importing to (SCRUM) and found whatever Issue matched number 379.

Workaround

There is no workaround. To avoid this, Issue creation and Issue update through a CSV file should be done separately to avoid this unwanted updates.