Basic concept: Patch collecting app. There are issuing entities in the system that issue patches. That model is "issues". In many cases, these are difficult to tell apart. So I want to add a similar issues functionality that would be self referential on the issues table.
When creating this, it should go both ways. If patch 1 looks like patch 5, then patch 5 looks like patch 1. Do I need to do double entry or would two indexes (issue_id, similar_issue_id) work?
This can often be many different issues, so would this make more sense to have issue_id and then have similar_issues_ids as a JSON or other multiple value store?