Duplicate rows can end up stuck in prearchive table

Likely due to some race condition, you can end up with multiple rows in the xdat_search.prearchive table for a single directory on the filesystem. When this happens, you cannot do anything with the data in the prearchive - everything complains about the duplicate rows.

Dropping the table and restarting tomcat will fix the issue by forcing a manual rebuild of the prearchive, but this is both time-consuming and requires downtime.

I notice there is no primary key nor uniqueness condition on the xdat_search.prearchive table, perhaps that'd fix the issue?