Splitting a database of indexed files

DEVONthink
1

I have a dozen folders (totaling about 68 GB) on my internal SSD that I have indexed in my Genealogy database. I would like to split off the Books folder (about 30 GB of PDFs) into its own database, still as indexed files. This database is not synced to any other device and the files are not synced to iCloud, DropBox, or anywhere else. I’ve read the relevant sections of the manual and searched the forums but I’m still uncertain as to the exact procedure.

One of the posts I read, if I understood it correctly, suggested exporting (Export Files and Folders) the group(s) to be split off, then indexing them into a new database. I’m not sure what the export accomplishes, since the files and folders are already external to the database; perhaps the procedure was meant to apply to imported rather than indexed files. What effect if any does this export have on the existing folders?

Can I simply create a new, empty database, and drag the Books group into it? Alternatively, can I just index the Books folder into the new database directly? Will my tags and meta data transfer with the files?

Finally, how do I then remove the Books group from the original database, without affecting the files on disk?

2

As always - make a backup first.

You are correct with your first scenario:

  • IMPORTANT: All files are indexed
  • Create a new books database in DEVONthink
  • Move the groups/items inside of DEVONthink into the newly created database
    • this will preserve your itemlinks

The result will be that nothing changes on the finder level all folders and files stay where they are.

I DO NOT RECOMMEND the following:

If you go with your second option (create a new db and just index the Books folder) the following consequences will be

  • you lose your existing itemlinks
  • if you have folders/files indexed into two DEVONthink database this can create problems when changing files in one database
3

@manfred9 is correct re:backups, not matter if you’re using DEVONthink or not.

I would not recommend moving 30GB of documents between databases in one move.

Is the Books folder indexed separately or is it a child of a parent folder?

4

I don’t think I have any item links, unless that includes annotation files (only half a dozen of those).

I assume by “move the group… inside of Devonthink” you mean using the Move To command? Does that also remove the files from the original database?

5

What would be a safe limit? I will need to break them down into subfolders.

It’s indexed separately.

6

You can use “Move to” command or drag and drop (without pressing any other keys).

In both cases it is a move and not a copy.

7

Annotations are part of the PDF document so there’s no worry about losing them on transfer.

How many documents do you have?

8

The Books folder has about 880 documents.

I was referring to annotation files (not annotations) – I suppose the link to the parent document would be an “item link”. Presumably these would be preserved on moving.

9

Yes, the backlink is an item link.

Personally, I would make a new group in the new database then sort the group in the source database by size. I would then grab a handful of PDFs in the source, Control-click, and select Move to > … in the context menu, choosing the new group in the destination database. After they transfer, grab another handful and press Option-Command-T to move them to the same location you previously chose.

Remember: This is not doing a simple file copy. It is also having to index metadata and contents of the files in the new database. 30GB for ~900 files means you have some very weighty documents, i.e., lots of text. So plenty of processing going on.

Also, I recommend you’ll want to make sure the Files > General > Annotations > Move annotations automatically setting is enabled. This way, you move the PDF document and the annotation file “follows” it to the destination database without you having to do it manually.

10

One approach I have used when trying to reconstruct large databases:

Back up

Duplicate the source DB.

If there are two folders in the source DB, let’s say they are named A and B:

Delete folder A in one of the DBs.

Delete folder B in the other.

11

This is not the correct thing to do. You should not duplicate database files in the Finder as the internal ID for the database is retained. You cannot have two working databases with the same UUID.

12

Noted. A technical issue that I was unaware of. Thank you.

13

You’re welcome.