Okay…, one more experiment failed.
First, I copied the digiKam files folder as a back-up.
I started digiKam, put it in Albums View, and clicked the very top “Albums” so nothing was selected.
I opened Tag Manager; and, I hi-lited the top of the tag tree “Tags”.
I then clicked Sync/Export (which is somewhat misleading, as I don’t see any way to export anything).
In the Sync/Export menu, I clicked “Wipe All Tags from Database ONLY”.
In a matter of seconds, the Tag/Categories pane went blank and the Tag Manager disappeared.
I waited a bit and then closed digiKam.
I waited a bit and restarted the machine.
I waited a bit and started digiKam.
I opened the Tag Manager and then the Sync/Export menu and clicked “Read Tags from Image”.
It told me that could take a while and I clicked to go ahead.
It did take a while.
I went to bed and came back a few hours later and it was done.
Results:
My tag tree was mostly there.
In the main parent tags, there were a few errant sub-tagged images listed as parent tags, not many, maybe half-a-dozen; in most of these cases, it would not be all of the images within the particular sub-tag group, but maybe three of several.
I did not meticulously explore any deeper than the main parent tag structure; I have no idea how many misplaced tags may have been scattered where they did not belong.
All of my tag icons were gone.
Now for the acid test:
I went to Tag View and tested some of the tags that I knew had been displaying incomplete information and missing thumbnails = FAIL FAIL FAIL
After all of this, I hadn’t gained a thing.
All of the corrupted data had obviously remained while the tags were wiped and digiKam just re-associated the re-enlisted tags back with the corrupted data — it did not purge itself of the corrupted data as I had hoped.
Back to the drawing board…
Before I performed my lengthy experiment described above, I found another distressing problem that may or may not be related to my big problem.
I noticed some antique and classic car images were also tagged as “Sawmill”
I know for absolute fact that I did not mistakenly apply the Sawmill tag to these images; also, the affected images were scattered as to when they had been inducted into digiKam and it would be a huge coincidence if I had made the same mistake with the Sawmill tag on all of those occasions.
I noticed the number of images tagged with Sawmill was many more than I thought there should be.
So, I checked the Sawmill tag in Tags View and I found many more wrongly tagged with the Sawmill tag than there were correctly tagged.
I had images of people, fire-trucks, Poland-China Hogs, 1972 Plymouth Barricudas, Minneapolis Moline tractors, Shetland ponies, helicopters, and more, all showing as tagged with the Sawmill tag; no way could I have made that many mistakes.
I don’t yet know if this discrepancy is restricted to only the Sawmill tag, or if I am going to start finding many more other wrongly associated tags…; I suspect and expect the later.
I have spent countless hours meticulously building my tag tree and carefully tagging my images and now my sole reason for doing so is self-destructing.
I fear it may be years before my bug report yields a fix or a solution.
I have considered deleting the database and starting all over anew; but I have very little hope that digiKam won’t somehow reconnect itself to the corrupted information, hidden somewhere within the bowels of my registry, and me be back in the same place I am after all the work, or just wait until I get everything to working good and fall to pieces again.
Thanks for everyone’s patience and suggestions thus far.
Sorry for so long and so much to digest in a single post; I tried to make this in two separate posts but it would not allow it.
Update 11-Dec-2018: After meticulously finding and removing all the wrongly applied “Sawmill” tags from random images scattered throughout my collection, this morning I found a few more totally unrelated files carrying the Sawmill tag.
I had hopes that migrating the database to a different HDD would possibly solve the problem; however, after a very lengthy fight to get it done, nothing was gained; and, if anything, things were worse.
I have given up on any means of fixing my corrupted database; and, one-by-one, folder-by-folder, I am building a new database.
I am going to be hugely frustrated if my problem shows back up in this entirely new configuration.