Restructuring a Database (Advice on)

Any advice on how to restructure a database or a preferred folder organizational method to reduce clutter?

I began using devonthink when I wrote my dissertation. It worked great and I can’t imagine going through that process without DT. I never settled on a single method for folder arrangement. The most helpful introduction was Rachel’s “DevonThink and History Research.” It refers to an earlier version of DT, but the methodology is sound. Like many on this forum, I have experimented with different organizational methods over the years. As a result I don’t have much of a discernible logical structure left in terms of folder arrangement, and an ever expanding database.

I can guess that @BLUEFROG will say there is no best way about this, and that it’s up to the user. Which I agree it is. Thank you Jim for your tremendous support over the years.

What I’m wondering is if anyone has encountered this same conundrum? If they went through a process of reorganizing a database after a major project? And how they approached it?

1 Like

You’re welcome :heart: :slight_smile: and you’re correct that my initial response is indeed that. My own own riff on the old adage: Some shoes fit some people; but no shoe fits everyone :wink:

That being said, I have used DEVONthink daily for almost 12 years, and do that vast majority in a DEVONthink database (with tons of ancillary databases for resources, testing, etc.). That main database is today called DEVONtech Support_r4, so I think you can guess it’s not my first iteration :smiley:

My approach has been periodically to examine how I use a database.

  • What do I use often?
  • What have I put in it but really never used?
  • What other approaches have I been considering for organization (but this is rarely looking at other systems, like PARA, Johnny Decimal, etc. except perhaps in a long-distance way)?

I create a new database – r5 (so brilliantly named - lol) – and copy over oft-used items. I then work in the new database but leave the old one open for a period of time as I transition to it. After that, I close the old one but have it available to cherry-pick from as needed.

The main point is: I examine how I work and think, not trying to reproduce other peoples’ approaches.


My preferred organization method is tags instead of groups (folders)
I prefer tags because multiple tags can be assigned to a note
which even supports different structures at the same time

I reflect hierarchy in my tag-names; for example
Budget-Housing, Budget-HousingRent, Budget-HousingUtilities, …

My top level tags are Type-aaaaaaaaa
which drives additional tags
For example Type-Receipt items also have tags for Vendor and Budget-Category
I use an applescript to assist with tag assignment

For restructuring; obsolete tagnames are prefixed with an x

1 Like

I’ve been reading about the PARA system, which suggests grouping material based in four categories: projects, areas, resources, and archives. I’m not sure I’m sold on it yet, but one of the key insights resonated with me.

File things where they will be actionable the soonest. Don’t put “windmills” and “solar cells” together under alternative energy. Put solar cells under the “new roof this summer?” project, and windmills with the “alternative energy” resources. (Or vice versa if “energy company wants to lease land for a wind farm” is a project you have.)

More generally, organize material according to when and how you want to find it again.

The thing that helps me the most is that I have a recurring task for myself to review 5 items in DEVONthink every day (I also review 3 older references in Zotero each day doing the same thing)–I open the item, make sure the name, tags, and other metadata are right and in a standard format, and then that it’s filed where I want it, replicated and duplicated into other groups or databases, etc. as needed.

5 a day is doable for me–but often I get on a roll and do a more. It helps me get through a lot of items as I re-structure and clean up.


There’s another alternative, which may sound radical at first.


Leave your existing database as it is, and start a new one for your next project. Trust the existing structure and DT’s search tools to find any old material you need as you need it. Move that material (and only that material) into whatever your new structure is.

The idea being that the benefits of restructuring aren’t enough to justify the time invested.


Indeed, this may be the ideal approach in certain cases.


I agree with everyone here but…
Whenever I’ve had to take over a big project from another editor, I’ve developed a strategy that would work in DevonThink. When you’ve been given what could be either a chaos nest or uniquely organized circus some extreme measures need to be taken.
I would do this:
Put all the files in an enclosing folder called “old structure” or “Project v1”
Then make a ton of smart rules to make lists of things by name or type or size or date. Keep making smart lists.
DTs replicate feature is amazing at this point because you can start replicating files from the smart lists into the new folder structure.
I’ve usually do this slowly while playing archeologist on the previous set up, watching footage and old edits.
Somewhere in there you’ve learned what you’ve got, what’s been done and where it could go.

1 Like

This may seem like an exceedingly stupid question, but I quite like this idea so I want to be clear on the following before attempting it: how do you decide which 5 items to review? I.e., what is your process for identifying items that need this review as opposed to the ones already reviewed so as not to update processed items but leave others down the list alone. I assume it’s systematic in some way, like, perhaps, a certain folder, or a certain tag or lack of a tag that then gets changed upon review, etc., but is that assumption correct?

1 Like

My reviewing process involves pulling a random note, which has not been reviewed in the last X days, in one of my databases. I use a special custom metadata field to record date of the last review, and a script to bring to me that random note. The script is not a necessity; a smart group is good enough for the job. I use a script so I don’t get intimidated by an endless list of notes waiting for review.


By default I dump everything into the Global Inbox, or the root level of the database.

Then I work through the Inbox to get to inbox zero–and the root level of a database until everything is filed into a group (this stems from habits learned using ProDOS 8 on my Apple II when younger–never keep anything in the root prefix!).

OR, if reorganizing a group structure, I rename the old group OLD Group and the new group Group and then work through OLD Group until it’s empty.


Last year, I came upon the Johnny.Decimal system, and this year I used it to revamp a collection that I have been using and continue to expand over time. It’s perhaps best described as a hyper-organized folder system. I have been happy with the end result.

And, as it turns out, in February someone posted a question at the J.D forum about using it with DevonThink.

I prefer tags instead of folders
and words instead of numbers
I reflect hierarchy in the tagnames


I’m still relatively new, but the one thing I’ve learned as I’ve tried to figure out how to organize is that whatever works for me today is likely not going to work for me tomorrow and so this is going to be a constant work-in-progress. I also recognize that I come from an erra when folders were the only way to organize anything, none of these tags or even longer file names like they have today. :slight_smile: A lot has changed since then of course, but some old patterns are just really really hard to break, at least for me. Given all this, I default to folder organization, but I"m trying to embrace tags more while resisting the temptation to over tag everything. The other thing that has been helpful for me is that when I file something, I replicate it to other places where I think I might need it in the future. Yes this is more work up-front, but it also gives me an extra moment to question myself about whether I’m putting it in the right place to begin with. For example, if I file something and then replicate it to three other places, going through that extra work gives me an opportunity to realize that I’m probably not filing it where I should be. My hope is that in time, I’ll worry less about finding things and pay more attention to how I actually go about finding them and then I can adjust from there.

1 Like

This is, in my opinion, the primary strength of tags over folders. It’s easier to modify the tags of a document than to deal with a scattered few of replicants. Merging tags is straightforward and effective maintenance. The fact that we have a tendency to over-tag things illustrates the flexibility of tags.

Tags are also more search-friendly. It is trivial to search for all items tagged apple, banana and candy. Finding all items with a replicant in each of three folders, however, can be a bit more tricky.

A stark contrast is in the decimal system used by @Yclipse , which is folder-based, single-threaded, and optimized for manual retrieval of physical items. Decimal systems are fundamentally based on the premise that you can determine whether something (books, in the case of a DDC library) suits your needs by its name/appearance alone. They do not take advantage of modern computers’ capabilities to process and combine multiple threads of information.


Agreed but I believe this is as much a weakness as a strength and calls for a bit more nuanced approach at least with critical data.

As an example I use DT3 as my primary respository of documents in a professional consulting practice. This means:

  • First and foremost, I need an accurate long-term archive of all documents associated with a given case. This is mission critical

  • Optionally it may be helpful to organize or identify documents in other ways- work in progress, background academic articles, telephone/Zoom notes, etc.

The ease of adding tags also means there is a risk of inadvertently deleting a tag. That could be catastrophic in the first scenario i.e. if I lose the association of a speicific document with a given case.

Thus in my main DT3 database, each “Case” gets a Group - since it is much harder to accidentally change a document’s Group than to delete a tag.

Additonal metadata for the document then often is best handled via tags.

I have to say that my opinion is to let DEVONthink 3 do it all. Obviously though BLUEFROG is right, as you guessed! I do make folders of obvious things now on import. Collections of papers on a very narrow topic for example. Even that is matter of opinion really though. In my case say ‘climate science’ merges with several other topics now, to my dismay in fact.
I just have a database for each part of the year, again no clear set boundary as to when I move to a new one and/or reimport the old files into that. Generally that pattern cleaves closely to my actual thinking and inquiry.

I basically use a pretty flat system, most stuff is a the top level with a few, now mostly historical, folders.
For example I have a paper which I copied using screenshots, it was very hard to download and inaccessible, so that sequence of screenshots has its own folder, I don’t hardly look at it nowadays anyway.

Then I use the search and smart rule functions on the flat organization. For example I have a smart rule that picks out my own notes an a particular topic, and indeed all my own notes, which include ones in “Finder”. Honestly I recommend trying it, it is a bit nerve wracking at first, but it worked for me and more like my own mind works. In fact it seems more like the old paper mountains where one, somehow, knew where things were all the same?
I really think over using folders and recreating Finder systems is not using DEVONthink 3 to its full capacity.


The emboldened phrase is a critical bit of information to read and really mull over.


Strangely I do what you do. Open a new one and kind of transition over a period of time and there is a kind of sorting process I can’t define. I appreciate it might be different is one has other kinds of data too.

1 Like

With tags, this is not a headache at all. Say you have just acquired a bunch of documents about climate change, natural disasters and climate change awareness. You just do the simple thing – to assign the three tags to each of the documents – while maintaining a flat system of organization, in the sense that all your tags are at the root level (that is, not nested).

At some point you may find quite a few tags with the name climate change xxxx. Then you can assign the tag climate change to each of these tags, and still maintain a flat system.

This method does not actually resemble how things work in my brain. I tend to think in chaotic, erratic, unorganized ways. That is the reason I seek help from PKM software in the first place. What the aforementioned method brings me is the ability to navigate related topics and documents seamlessly, in the sense that I can repeatedly Reveal Tags without having to build a search/smart group or dig through a tree of folders. I believe this saves me more time and energy than those spent on organizing my tags.

1 Like