I am fairly new to DT, been using it about six months. I am an elementary school principal, so I have lots of stuff to put into DT. When I first set it up, I started with five databases (Memos, Misc, Parents, Staff, and Students). I was recently trying to create a replicant and discovered that you cannot create one in a different database. This is problematic for the parent, staff, and student databases because saving an email in a parent’s group it would be useful to have a replicant in the teacher’s and student’s group as well. Naturally I am thinking about combining the three databases into one and then just using groups. Then I thought, gee maybe I should just EVERYTHING in one database. I am curious as to what your thoughts are on doing this?
Because of what you say you want to do and assuming you are not going to have millions of documents… seems like One database way to go unless for reasons, say security and access, you really need more.
I use a single database,
and would use tags for those categories
for example Type-Memos, Type-Misc, Type-Parents, Type-Staff, …
Note; In Devonthink, tags are also listed as special groups
Not everyone can shift to a Tags only or Groups only set up instantly, I couldn’t.
What I’ve done over the last three years is slowly transition from one giant database to 7 ongoing databases. Some are very tag heavy and some rely on almost entirely groups. A lot of that transition is figuring out efficient tag structures and group organization.
It’s evolutionary and, in the long run, revolutionary.
Could you share how you think about what should be a tag versus a group for you? I’ve stuck with groups for now but don’t have a firm understanding or examples of the differences
I use one major database and several content specific databases. The content specific ones don’t overlap and it’s easier to keep those very specific. Receipts is one example. The major is a reference database that I dump everything else into.
Tagging, especially once you’ve discovered nested tags, can get powerful.
Genres, styles, anything that’s already hierarchical or can be put into some hierarchy will suit tags very well. Tags are also great for things that have multiple attributes.
Tags work well if you’re a planning ahead before chaos type of person.
You create your tags, make all your smart lists, and then as the docs pour in tagging is a fast way to “send” them to the proper places.
Example: Episodic television production or post production
Tags(nowhere near a complete list): episode number, revision numbers, type of doc (call sheet, script, screening notes), department (editorial, special effects, sound), priority
Docs start to come in fast even before it’s time to roll camera so throwing them into the inbox and tagging them takes less time than replicating to the various folders. I find in those situations with tags there’s less of a chance of something getting missed.
I find there’s much potential for overlap
which is lost when data is siloed in databases (or groups)
You mentioned receipts which likewise I store in Devonthink
They’re great for generating expense/budget reports
An example is a purchase from Amazon
I assign a delivery due-date tag and the record becomes part of my task management process
For my home inventory process, I assign a tag
I see tags and groups as some kind of orthogonal stuff. You can organize your files by groups in a file-like hierarchy, and then cross over that with tags.
For example, I have some collections of Jules Verne books organized by folders, each folder (group) has the whole collection, eg. ORBIS 1988, RBA 2003, RBA 2008, Jubera, Roig, Sopena, Trilla. Each of those is a folder inside of my Jules Verne folder (or group, I’m an old guy used to folder terminology).
But I want to know, for example, in what book appears, say, women. Then I tag each book that has at least one woman in the story. Or has engravings inside, tagging each one with that word.
When I need to know how many Verne books women in them have, I search or go to the tag tree, and I have all the books I’ve tagged. When I want to know how many books one collection has, then I go to the folder structure (or search for the folder name).
Without tags, I could have built another folder structure and then use replicants to have in each folder all books with women, or with engravings…
Or you can go other way around: put all the women-in-books in a folder and tag the collections. Luckly, with DT the limit is not the program, the limit is you.
@DTLow Apologies for bumping this old thread, but I’m also currently moving from Evernote to DEVONthink and am wrestling with this problem. I see the clear benefit of having only one database, but doesn’t it complicate syncing with mobile, in that you might have to download a huge database, instead of selectively downloading the databases that you need on the move?
What has your experience been with this? Thanks!
A full sync of my Devonthink data works for me
If I need selective sync, I would have multiple databases
edit; btw I’m not experiencing any syncing complications
OK, good to know it’s not too cumbersome, thanks.
I have two primary databases.
I really, really like to keep my work and personal things separate. I have a work laptop and I don’t want any personal information on it for more than a day. I use the Global Inbox 99% of the time when adding something to DT. If I’m working and think “Oh, here’s a thought that I want to capture for my personal database!” I’ll write it in the Global Inbox and then after I sync I can just move it to where it needs to go in my personal database on my phone. On my phone I sync both databases.
I create a new database every year for my work and personal databases (unimaginatively titled “Work 2024”) so that things don’t get too wild.
Oh, and I only use Bonjour because it’s fast as hell and gives so many less errors than anything else, in my experience. Also, as far as I’m aware, it’s one of the most secure options since none of you data sits on any servers.
What things belong together?
What information do you need to access when?
The chance that I’ll need to access financial records from my phone is much smaller than the chance that my phone will be lost or stolen. So my phone doesn’t even know my financial database exists.
My fiction and non-fiction writing use completely different source materials, so I have separate resource archives for both. On the other hand, I might need to access either while traveling, so my iPad has access to both.
My point being that this is a “how long is a piece of string” kind of question. Thinking about what you actually need is more important than what other people do.
@kewms Thanks for that. I agree with your points but my question was really a lot more basic, so that my options are clear: if I have a huge, single database, can a phone handle syncing the entire thing well, or is this, simply because it can’t, another argument in favour of multiple databases?
The size is relevant only for the first sync, i.e. when you initially upload everything to your phone. After that, you’ll only have small data blocks to sync (probably – unless you store 4-hour movies in 8K or use other huge data blobs regularly).
OTOH, having a single database might complicate things logically, i.e. if you’re putting unrelated documents in it.
See the Getting Started > Building Your Database section of the built-in Help and manual for insight and our recommendations.
Thanks. I’ve read the recommendations (e.g., “DEVONthink’s performance can suffer if a single database grows too large. That applies to syncing performance too.”) but also noticed that some longtime, expert users of DEVONthink don’t follow them, so I was hoping to get some clarity on that, which @DTLow’s terse answer gave me.
I think we’d absolutely be classified as “longtime, expert users” 
PS: I have four main support databases (two of which are open 99% of the time), two personal, one financial. Beyond that I have an ever-expanding and contracting pool of test databases for support.
I think we’d absolutely be classified as “longtime, expert users”
Of course, but there doesn’t seem to be universal agreement on this matter, among experts (notice I said “some longtime, expert users…”)
My question was aimed simply at clarifying this for my own sake, not at starting an argument 
Thanks for letting me know how you’ve set things up, it helps.