I’m not a lawyer but work in the grey area between lawyer and surveyor. I use DT3 (and Scrivener) for my law library. I keep pdfs in Dropbox with a (parent) folder for each topic, for example LAW (general), LAW- Cases, Law - Books. The folders are each indexed to DT3. By using Dropbox as the intermediary, I can then use DT3 on another machine without any of the transfer issues I was encountering when using the DT3 database on one machine as the go-between for the other.
Since I started indexing folders from Dropbox I haven’t encountered the same issue as the op but I did when folders were in Finder.
Whether it is just as ‘safe’ to keep the folders in Dropbox as in Finder, the other day all the pdfs in one folder in Dropbox went missing - I had to restore from Time Machine. That made me think it might not be such a good idea to keep everything in Dropbox.
I have considered using DT3 in the way you describe. Currently I have a DT3 database for all calls, notes, copy/paste. Everything goes into its inbox and I have smart groups for each client, having tagged each record with my abbreviation for the client’s work. As and when I have completed a job, I can print everything to pdf and keep the information in a central folder and sub-folders I have in Documents. (For contact info, and emails I am using Clio.)
“I understand that if I just created a new DB called “Cases” and then indexed cases from the HDD added as “groups” to that DB I would not have the issue. BUT, then I would have a huge database of tons of cases that has to be open all the time, even when I just want to work on one small item.”
In Dropbox, I have a folder “LAW - Cases” which contains >5000 law reports and digests. I have linked that folder to my DT3 Law database. If I were to keep my work also in that DT3 Law database then by using tags I could tag the case law to a smart group for the client so that anything pertaining to that client would be easy to isolate.