Thanks, Bob!
The question is
- did .72 generate a correct library (about 105.000 entries), and .73 has a problem of correctly reading the library (showed up about 100.000 entries), or
- did .72 generate a library file with flaws, i.e. records that were damaged in some way, that itself could read but .73 failed to do so.
To clarify that I had to generate a clean .73 version of the library. I imported manually (no auto-import) the 26 directories (A-Z) and .73 found 109.219 records.
My conclusion:
1. the .72 library files have some incorrect (damaged) entries
2. the .72 version managed to lose some 4000 records (.72 - abt. 105.000, .73 takeover - abt. 100.000, .73 clean, new: 109.219)
3. it is not clear the the .73 version does not have the same problem when manipulating library entries.
What did I to the .72 library (and the versions before): When entering or updating albums I did one of two things:
- when I simply added some albums to the directory of the artist, I had MC import either the album (specifying the album directory, 1 or 2 albums) or the artist directory (more that 2 albums).
- when I had multiple changes to the artist directory (deleting albums, modifying album names (directory), filenames, or exchanging mp3 for flac, I deleted the artist from the library and (re)-imported the artist directory.
(and not to forget I switched OS versions in several steps from 10.14.x to 10.15, starting with 10.15.3 and now 10.15.4)
Somehow during the whole process the library file got corrupted...
Size of the backup files: .73 - 4 MB. .72 - 4.1 MB (??). -- why is the .73 backup with more library entries smaller?
(I'd think that the format of some library entries changed...?)
So far my report.
What will I do differently now: I keep a close eye on the totals in the library when I modify something in my collection.
Maybe there is a tool to validate directory structure... I would apply it...
Is there?