interaction between databases - Possible? / HowTo?


I’m sure this subject will already be touched upon, but could not rapidly find anything relevant.
Maybe I’m overlooking things, if so, please point me to where I can find the information…

What is the inter-database capability of DTPO, if any?
Are there any future plans to develop this?

I’m thinking of:

  • granular / parial sync between DB’s
  • granular / partial sync between groups / subgroups
  • replicate / duplicate / copy between DB’s
  • link between DB items works, so this is great!


Why would you want to “synchronize” two databases on the same machine?

You could, of course, index the same external folders to multiple databases.

I wouldn’t.
I would, if possible, sync subsets.

I KNOW that there is a lot that I still DON’T know in / about DTPO.

The idea stems from the following situation:
Company (1980): 
Folks, today we announce our plans on our new filing strategy. As you all know and have experienced, often times documents get lost, are misplaced, misfiled.
After long and exhaustive debating we think we can now come up with THE solution.

We are going to file documents on a per country basis.
Everybody: Wow! What a good idea! 

After 3 months:
Jim, where's the document X? Ah, I filed it under it's country: America
Jill: Jim, I can't find it..
5 days of searching later..
Jill: I found it! It's in the United States file..
I'm going to prevent this from happening again..

3 months later again:
Peter: Jim where's the document X?
Jim: I seem to remember it was filed under United States.
Peter: It's neither under United States nor under America.
Peter: Ahah! I found yet another map named U.S. (or Verenigde Staten or whatever is your language..)

Not sure, but I think this is / can be expressed by a one to many or many to many relation issue.

I (try) to solve this by creating replica’s of items or groups.
So far I have only used this within my main DB.
Did some experimenting today, and item-link spans across DB’s, which is perfect.

If I could extend this to groups / databases this would extend capabilities by a lot. ( At least so I think with my limited knowledge of today)

On the practical side:
Also keeping DTTG in mind, but not only, I would keep certain things separated, and also at least ‘consider’ sync bandwidth, and unnecessary loss of diskspace & proliferation.
E.g. all manuals (pdf) go into DB_manuals.
In short all the possible reasons why you would keep stuff in a separate DB.

(from a philosophical point there are many other subjects that come into play of course)


So, what specific thing are you looking to do that, e.g., tagging and Search and Smart Groups and See Also, etc., cannot do?

For example, as far as I know, correct me if I’m wrong, a replica of e.g. a pdf, will keep only a single pdf in ‘the’ DT filesystem.

Not sure, and never looked into it neither, but I suppose / guess that if you sync a DB to

  • syncstore
  • other mac
    a ‘different’ underlying file structure is built, and kept in relative sync by syncing the overlaying DB.

I would very much like to be able to create cross DB replica’s.
At the moment I can use cross DB x-devonthink-item://xyz links.
I can use these in text, rtf, probably sheets (don’t use these much).

But I can’t replicate tags / groups…
So, I cannot maintain a similar group / folder / tag layout across DB’s. The way it appears on a users retina will differ…

I’m not sure if I convey my idea clearly?
Returning to the example:
If I have the liberty to ‘file’ documents relative to Automation into my Automation DB, but also can keep a replicated ‘group’ in my ‘main’ DB, I would always find my document if I search in the Automation DB, but I would also find it in my ‘Main’ DB.

Also, suppose I share a bookkeeping & physicality DB.
And I sync this with a mac that is mainly used to do bookkeeping & administration related tasks, then the ‘right’ documents will be added to this machine, while I will also know and be able to see in my main DB, that granularly synced items will be available (as a replica) on that specific machine.



You’re looking for a data structure that does not exist as specified (if I actually understand your specifications) in DEVONthink (or any of its competitors on OS X – as far I know).

Except you can do much that in a file system with a file system browser such as Finder, Path Finder, etc.

Perhaps you need to combine databases rather than deconstruct them? Or do more indexing?

In the past, more my windows era, I used Ecco Pro. I suppose you must have come across Ecco Pro?
The moment I started using it, was already long after it’s commercial availability.
Even today it’s still ‘kept alive’ by a couple of die-hards (in the positive sense) but not really ‘supported’.
Even at my time of using it, I never used all its ‘intricate’ ways of interacting.

Totally different, and yet there are a lot of similarities.

I indeed have been thinking about ‘better use one DB’ instead of several.
Mostly I like to keep pdf magazines in one DB. I don’t do it consistently yet.
But I fear this way the DB to quickly become too large / difficult if not impossible to handle with DTTG.

=> What other ways you see to ‘combine’ DB’s rather than ‘merging’ them?

=> More indexing…
I’ve cut my fingers and more than that by too much indexing.
That was the time when I was doubting whom I could not trust: myself or DTPO :wink:
(still have this at times, but DTPO always wins…)

I’ve e.g. more and more started to work with DTPO in a way that it is ‘my trusted filesystem’.
And I’d like to keep it like that, even learn how to get better at it.
But e.g. I find that when I start out with a RTF list, say a daily journal where I take notes, after some time, DT creates ‘copies’ of the file. I may even end up with a number of copies, which I then need to find out what is missing from which copy, and which copies I can trash…

Also there are quite some differences in the way file types are handled.
I for one do a lot of searching on the web. I then very much like to ‘select’ part of a page, just drag’n drop to ‘the other’ app, and have it readable, not copy formatting, keep links alive, have a link to the original in the result.
This is possible within certain limitations with multiple applications.
Scapple e.g. is very good at this, and keeps things simple.
Storing Scapple files within DTPO are easy to ‘look at’ but they are not searchable by DTPO…
Curio is wonderful at a lot of this, but also has its limitations. (store a mind map separately)


This is likely due to the Conflicts setting you have in DEVONthink’s Preferences > Sync.

No, cross-database Replicants are not possible.


Thanks for hopping in!
Any chance this gets on the future requested items list?

Preferences > sync makes sense…
I have it on Duplicate documents. It seems that sync happens about every 30 minutes then?

Use latest document, I gather this will just keep the original as long as the document is open in the editor?



It’s been requested many times. It’s a technical issue, not a philosophical one.

Sync triggers on a User set interval or from 30 seconds after a change to up to 4.5 minutes when idle, on Automatic.

Newest should preserve the last saved document. And yes, we still suggest doing manual Saves often.

OK - I have a very similar desire. Jim Neumann was pretty clear when he said “No, cross-database Replicants are not possible”. None the less I’m stating my situation and maybe one of you gurus will have a solution.
I have General Research Data that covers multiple areas in on DTPO database. I have another DTPO database that relates to specific projects that I’m engaged in. There are times when I have already have research results (in Research DB) that are related to a new project and I want those research results tied to the current project (Project DB). I don’t want to “move” the DTPO entry from one DB to the other. I can’t replicate the data, per Jim’s comment. and I don’t see a way to “copy” the DTPO entry. So, what can I do?
Thank you all for your help!

In your research data DB, tag the research related to the new project with a tag. Use the same tag in your “Projects” DB (or some other attribute that is available in the Search window, could even be belonging to a specific group)
In the sidebar of DEVONthink, create a new smart group. Set the criteria to show files tagged with that projects tag (or tagged with the project tag and has whatever the other attribute is e.g. “belongs to group named [new project]).

These sidebar smart groups will reveal files that fit the criteria from ALL open databases.