I am making a large archive of different aspects of work. Is there an easy way to group them without having to do it manually (thousands of documents)
Maybe I’m misunderstanding your question, but do you know how you want to group your documents? I’m afraid DT doesn’t know ![]()
But at least in theory, you can create rules… move all documents that meet this criterion into this group.
Thanks Frank. I’m coming into DEVONthink with just a huge number of documents (30k), which I want to sort into different groups (fiction, non-fiction, scripts, research, different titles, agents, etc.) It’s all Word docs - just different folders where they should live. Was wondering if there were shortcuts to doing this (doing it by hand will put me into my 80’s… I’ve been watching some films on youtube about tags and smart groups, etc. Need to learn more about rules. I think that’s the short cut I seek. I will continue to try to come up to speed. Appreciate your time!
If you already have the documents in DT but they aren’t yet grouped the way you’d like, smart groups might help.
You can create different smart groups, for example, all documents with property A, all with property B, and so on.
The advantage of this is that you don’t have to move any documents. You can leave everything as it is and still group them (virtually).
But others might have better ideas.
Ah, one more thing: DT has an excellent search function. Anything you set up using smart groups can also be done “manually” if and when needed.
Welcome @jeanstraussfilms
-
How were the documents stored in the Finder – loose or in folders?
-
What kind of metadata is attached to the files, e.g., were you using Finder tags?
-
You shouldn’t be imagining processing 30,000+ documents in one fell swoop. You are talking about a lot of processing and if you “set it and walk away” and something goes astray, you would come back and be frustrated unnecessarily. Small gains lead to more consistent and controllable successes.
Batch processing is the way to go here (or doing a blitz for a week - but that eats at your soul a bit so is only an option if you have to, or if you really like to make yourself suffer).
Once you’ve defined your criteria for each group, you can start thinking of smart ways to process files in batches. For example, let’s say “Scripts” is a group, and you know that they always contain a unique search term, e.g. “Script for…”. You can now set up a search for that term and move all the files that meet it into the folder. The trick is finding enough unique search terms to gather up a chunk of the files that will meet that criteria. Ideally, you want to get through the bulk of your files with clever searches like that. Maybe files detailing agents always have an email or postal address and other files don’t. Maybe you only ever create research files in January. It can take some time to figure out unique searches, but makes your task much easier since it allows you to process in bulk. You might also want to rename the files now they’re in their correct locations; rules are the way to go for that.
Eventually you’ll be left with a small group of files that didn’t meet any of your bulk criteria so need to be processed with ever smaller criteria or by hand. It’s much more manageable to look at that small group.
I’m guessing it’s taken you a lot longer than one year to accumulate so many files, so even with the help of tech it’s good to acknowledge it will take you some time to tidy them all up. DT search is really good so even while you’re doing your big tidy up in batches, you should still be able to find the things you need. And of course some people never organise their files into groups, they just rely on search. I’m not one of them; I’m a huge believer in the importance of a good filing system, but I’m saying it because not everyone agrees and that’s ok! Search is great!
For what it’s worth, I took a few days and did most my processing in one go. I just wanted to get it over with. But I’ve seen people just do 15 mins a day of filing (you’d maybe do a few searches a day and move your identified files to their new homes) and it can add up to a nice filing system very quickly.
Once you’ve got a decent chunk of files in their correct groups, DT has a classify function where it can suggest groups for you. It will need data to learn where things should be placed though so you need to “train” it by doing work at the beginning (I don’t use it much personally because I usually know where I want to put a file, but in my experience it is right about 90% of the time in my very structured database).
You can also use an LLM to do all this filing. I’m sure others can speak to that (I don’t use them so can’t recommend a process), but I will note that there’s bound to be a couple of unexpected anomalies (LLMs filing things in the wrong place) so you need to weigh up whether doing bulk processing with criteria you set is worth it for a better end result (more accuracy at the cost of more time). For me, I care a lot about my structure (I expect things to be in the folder I want them to be in) so it’s not worth it for me in my main database. But if it was a database where I didn’t really care that much if a few items ended up in different places, for the numbers you’re talking about I’d be inclined to ask the computer to do it.