I created a library with a few thousand content items. I populated this by dragging groups of files into an initial empty database. For each group of files I applied tags. I have no groups, and use the tag folders for navigating.
It occurred to me (sadly, only after I had done all this) that I might have been better off not importing the files, but rather indexing the content and leaving it in the file system. I realised that I could tag the database content using either approach.
So I have been wondering; is there a way of converting this database from an internal storage database to one that has indexed the external file content?
I can see how I could do part of this; I can export files and folders to the file system, resulting in a single folder with all of the content, then indexing that content from within a new Devonthink database. But I can’t see how I could preserve the tagging, which is very important as it is the basis of navigation in Devonthink.
Does anyone have thoughts about how I might do this? Thanks for any advice.
Why are you “better off” ?
Exporting and reimporting includes the tags. It does not preserve item links.
PS: This has been asked more than once over the years, including a few weeks ago with the instructions for switching but preserving the item links…
I should have said “better off for my individual use case” - it’s probably not how most people would work. Most of my work is with RDF knowledge graphs, and so in general I prefer distributed systems that allow me to keep my content where I created it, and having a management system to hold classification and links.
Thanks for the pointer - I will check that posting.
Interesting and thanks for sharing.
I don’t think that other article fits my need. I’m not worried about being able to index the exported content (I know that works well) and I don’t have any item links (I assume it means item-to-item links) so I’m not worried about preserving them. It’s the tags assigned to items that I’m concerned about. Or maybe I’ve misunderstood and tags are also links. If that’s the case it’s worth a try.
You asked:
Is that not explained by the linked post? (If you want the content indexed, there’s no need to first export it only to index it again.)
Regarding tags, BLUEFROG already said:
You might also want to look in the manual. Getting Started > Tagging > Finder Tags should clarify things. It would probably be good to read the whole tagging section.
This is interesting; it’s looking more promising (thanks, @BLUEFROG).
I exported the database of imported content. This created a file for each content item, plus a file called DEVONtech_storage. I created a new empty database (DB1) and imported the content and the tags were restored.
Then I created another new empty database (DB2) and populated it by indexing that Finder folder. This also created (I’m guessing) pointers to the content, and also the tags.
Finally, I created another database (DB3) and populated it by indexing that folder again, but first deleted the DEVONtech_storage file. I’d speculated that this file is where the tagging information is stored). Somewhat to my surprise, the resulting DB3 database in Devonthink has tagged content too.
So is the tagging in Devonthink available to all databases (even those not yet created) and if so how does Devonthink “know” what tags belong to what content? I don’t think that information is in the exported Finder content, so how is the tag restored?
Every database has its own independent set of tags. If you export a document with a tag and import into another database, the tag is added to the tags of the new database if it doesn’t already exist there. And the DEVONtech_Storage file carries metadata about the exported items.