When I import pdfs into my database (23.6 GB, 7200 documents, 5600 pdfs, 1,500,000 different words, 57,000,000 words altogether; Mac OS 10.6.8; DT ProOffice 2.3.5) it regularly happens that they get “orphanised” (Tools -> Check & Repair). It roughly affects one out of three or four. Sometimes, when I manually correct the name and put the orphan back where it belongs, remove the empty file and do Check & Repair again, the same file is again marked as orphaned!
While in the end I can always correct the orphans until Check & Repair finds no more errors, this behaviour is VERY annoying. I wonder if it affects other users as well (yet, I was not able to find similar posts) or if it has to do with my setup/database. However, remembering earlier posts by Christian, its size and word count should still not be critical.
Orphans occur when a file makes its way into the DEVONthink database package but the database metadata does not know about the file. So, could you go into detail how you are importing files? Why is it necessary to “manually correct the name”? Also, you say the files are “empty” – is there anything about the file that indicates it’s source?
Here is a common cause of orphan files: open a PDF in Skim to read and annotate. Skim creates a .skim file containing the annotations and puts that .skim file in the same folder as the PDF. If the PDF was inside a DEVONthink database to begin with, the .skim file will be placed in the database package but DEVONthink won’t know about it. Verify & Repair will then find an orphan. (It gets worse - sometimes the .skim file gets moved elsewhere in the database and Skim can no longer find the annotations. This is why I never import PDFs that have Skim annotations – I always index them.)
I use Drag and drop for importing. The names of orphans get truncated, so I have to readjust this. The original file is without content.
In my case, orphanisation happens after I have moved/replicated the files. I can see why that might pose problems then and now, but it happens far too often.
None of this is DEVONthink’s ordinary or normal behavior. Over here, Importing never changes names, moving and replicating never orphans files. I share your frustration - but these problems just don’t happen unless there’s something wrong with the database, the installation of DEVONthink, or the filesystem or maybe OS X. I suspect your database is corrupted in some manner. What happens if you do a clean install of DEVONthink and create a new test database?
Otherwise, the database behaves normal, I also regularly invoke Backup & Optimize. But thanks for the suggestion. I have to try this over the weekend.
Sorry for being imprecise. The files become empty after orphanisation, so I end up with a truncated orphanised file (in the automatically created group “Orphaned files”) and the original file in the correct location, but without content. I then delete the latter, rename the former and move it to the location/folder where it should be.
The problem seems to have vanished after I have moved my main DT database to another mobile hard drive. I suspect now that the problem has arisen from the file system: for compability reasons, the old hard drive was formated FAT, the new one is HFS+. Obviously, DT does not like to reside on FAT-drives.