Do you Import or Index?

I import data for processing, but it then goes to indexed folders for archival. I have various Smart Rules set up to handle most of that for me.

Only indexing. 7 databases on Mac, Bonjour sync to iPad and iPhone (but only a selection of the databases). No problems so far thanks to the according chapter in the manual.
I domit this way because I prefer to have my files not in the databases.

1 Like

Devonthink Pro is my digital archive. For archiving purposes, most of my content is imported. If I need to annotate or perform other “destructive” operations on documents in Devonthink Pro, I will immediately create a copy to do so. I will choose to import if the impact of the operation is clear and the security of the operation object is controllable.

Indexing is mainly for “dynamic” files. These documents have a high editing frequency and can be better processed in external editors. For example, I choose to index the Obsidian folder instead of importing it. Obsidian has a huge plugin library, numerous visualization options, flexibility, and speed. I will use this software to create a large number of interconnected notes, and link them to the content in Devonthink via Devonthink’s item link.

I am following the academic dynamics related to digitization of libraries, archives, and museums in order to have a better strategy for my personal information management. Devonthink Pro is the most satisfying software I have tried.

Regarding import or indexing, you can try to consider these factors:

  • Access frequency
  • Information importance
  • Priority
  • Information update frequency
  • Information tagging status(Metadata)
  • Information completeness
  • Information retrievability
  • Information complexity
  • Information content controllability:
  • Information relevance

    I hope this is helpful.

Did you know you don’t need to do this? If you’re annotating PDfs, you can export a clean version from DT at any time, and it would strip out the annotations. This means you only need to keep one version of a PDF.

The menu is: Data > Convert > to PDF without Annotations


I am late in reading this, but my response is that I have a Bookends database with 25,000 attachments (27000 references including ones without attachments) and it works fine. I have had speed issues, but those could be resolved by altering display settings. So if Bookends does what someone wants, I would not be deterred by the belief that databases must be kept under 10,000. Yes, it is not quite so quick with a lot of references, but performance changes very slowly as you add references: it does not fall off a cliff.) In addition, you can have multiple databases, and move references from one to another (I have archived roughly 10,000 references that I am unlikely to use again that way.) Though Bookends is snappier on an M2 with more memory, it was working acceptably on my 2018 Intel Mac Mini with 8GB of RAM.

Two merits of Bookends are that it is flexible, with an enormous number of options (including ones to manage attachments), and that it can be integrated with DevonThink in several ways, most of which are discussed on this forum. There are many other ways, but I am currently indexing my BE attachments folder in DTP3, so that when I import references using the built in command, and with a carefully chosen BE export format, those references (RTF files) contain wikilinks to the PDFs. I can then organise things in DTP, with one click taking me back to the reference in BE.