Scalability/Maximum Database Size of DevonThink

What’s the maximum number of documents DevonThink can handle? I’ve found several contradicting statements in the forum:

  • 200k documents / 300 million words (from Christian)
  • one user reported having 1.5 million documents in his database
  • another user claimed a noticable degree in performance after only 5000 pdf files.

What are the real limits? What are the consequences when one exceeds the limit? Are the limits per database or apply to the sum of all open databases? Does it make a difference wether the documents are imported or indexed? What other factors influence the maximum database size, e.g. does the document type matter (pdf vs. web archive vs. text)?

Regards,
Alex

The number of words applies to the sum of all open databases and is the most critical number. The number of items is only secondary and just a recommendation per database.

Therefore 5000 huge PDF documents (e.g. books) can easily exceed the recommended number of words. But it might be also a virtual memory issue (depending on the amount of RAM and number of applications running concurrently).

Finally, importing/indexing doesn’t make a difference and usually the document type doesn’t matter.

Thanks for this detailed answer!

So the limit is about 300 million words (as long as enough virtual memory is available). For clarification, does this refer to the number of unique words or total number of words?

The total number.

Taking this issue to an email archive. I am archiving all emails from prior years. The total number of emails is about 70,000, and the total size of the DB is 11.9 gb. I don’t know how many pages of attachments thre might be.

It takes several minutes to even load.

I’m assuming that the DB includes all attachments, and that they need to be taken into consideration. When I imported I did so from Apple Mail.

Is this assumption correct?

No, attachments are not indexed, so do not contribute to the total word count.