Hindsight is always 20/20, so I have likely made some fundamental errors in my approach. Hopefully there is a way I can get myself back on track.
I currently have 2 RM databases - one that I use as my “Primary-Work” where I am still validating quite a bit of the data, and where I do primary work. The second was on I treeshare with Ancestry and contains data I am confident enough with to share publicly. I hoped to use the Ancestry RM file to fill in gaps and help with stonewalls.
That means I currently have 2 files that are being added to - one via treeshare sync, the other that I have some manual additions to.
Right now I am currently trying to reconcile the differences and trying to use RM9 Compare Files. My issue is that when I selectively copied the records from my Primary to my Treeshare RM file the record ID’s were not preserved, so when I run the file compare virtually every record is different, with a huge number being the person ID which appended to the primary database entry and have a % match that is 90% plus.
I don’t mind manually reconciling actual differences between the 2 files, but now I have several thousand of names that really compare no differences
I think what you are asking for is finer resolution (better discrimination) of the differences between the facts and sources of persons in the two databases that you are confident are the same person. Had the Record Numbers carried over, you would be more confident? If that is all you want, I would have advised you to export the people of interest to GEDCOM, imported that to the new, empty database with the option “Preserve record numbers” and then uploaded to TreeShare. Instead, you likely used drag’n’drop which has no such option when the target database is empty and so they were renumbered.
Also, there is a hidden program-generated “Universal” ID assigned to each person which would have been carried over by GEDCOM but I’m unsure if it is via drag’n’drop. In RM7, that would flag them in File Compare as the same person; I haven’t checked in RM9. That’s the hindsight.
Going forward, it’s possible to update the hidden UID’s of the TreeShared database to match those of the working database at the risk of some ambiguity - maybe restricted to only those with matching primary names and birth date or death date - using SQLite directly on the database. Revising record numbers would be aided by having matching UIDs but is even more complex because so many tables link to the person via their RIN. I don’t know what else could be done to facilitate the comparisons; there’s nothing in RM.
Yes, I agree, I could/should have created the file differently. My concern is not that the record numbers are different, or any degree of confidence, but that this difference is being highlighted on the file compare. The result is that I cannot, without going record by record, tell that this difference is factual (i.e. a difference in a birth date) or, for lack of a better word, arbitrary (the numbering system internal to RM). Since the latter is not anything I can change I have no way to “fix” any of the discrepancies and they will constantly appear in the thousands of “differences” between the 2 files which makes the file compare utility fairly useless.
I agree that since the RIN is the key to all the associations, it will be unlikely that there is a solution to my problem other than removing my ancestry tree and starting over.