How to start building a database?

I am curious whether DT veterans started their databases by first designing an empty hierarchy and then filling it in by importing files or whether they simply started by importing copious amounts of files and then imposing a hierarchy by moving things around in their database.

Also, I intend to significantly reduce my Documents folder by simply creating three databases (Teaching, Research, and Other). Virtually every file in my Docs folder would be included in one of these large databases. I am still confused—and yes I went through the tutorial and Help----about whether DT Pro imports everything (except mp3 and images) and whether it is “SAFE” to delete any files left in the Documents Folder (ones already imported).

Any help or advice greatly appreciated. Refer me to the FAQ, Tutorial, or HELP if these concerns are addressed please.

My advice is NOT to create an “empty hierarchy.” If the materials in your Documents folder are in an arrangement that you understand, just replicate that. Or create three major folders: Teaching, Research, and Other. If there’s overlap between those three, you’ll get better results searching inside one database than in three. As you gather new data, it’s easier to work with a single database, too.

It probably is safe to delect documents, once they are in DT, but only if you make nightly backups of the DT files. On the theory that a belt and suspenders may both fail, I tend not to throw away original files. If storage is an issue, you could burn them to CDs or store them on an external hard drive.

What I did is to start with an empty database and add documents one by one. For each document I either find a way in the database or open a new folder. As it grows I feel that some issues need subfolders and this goes on. It takes some time but it was very enjoyable for me to go over what I have and what I want to keep for the future. Also I now know my database structure by heart.

I agree with Will (howarth) and also went through a similar development process as physicist jedi (cool moniker) in the creation of my db. I’d like to add that having all three major folders or areas in one db is definitely the way to go, for searching and just to make life easier. You can create separate windows and organize each area separately as you like, and it’s very much like having separate dbs that way, but with the advantage that they are all housed within the same overarching db structure. I find this to be highly effective both for being able to separate information out to organize it and work on it separately, while having everything available for doing searches and using the ‘AI’ functions DT has to offer.

Just my two cents!


I started with the creation of an empty hierarchy based on my ongoing research. It was basically the same as that in my file cabinet, which was duplicated by my hard disk organization.

Then I created a group called “Drop Box” which was where I imported stuff, massaged it, then sent it to the appropriate group. I might add that my research involves a number of seemingly unrelated topics. I use DTPro’s ability to label subjects by color to help the organization. (I just wish there were MORE colors to help.)

As my research has grown, I have made many, many more subfolders for two purposes: organization, and somewhere in the program docs it says it searches the database much faster with many folders rather than loose documents.

However, with all this, my main DTP database view takes up less than one page.

I hope this gives you a hint about one way to figure out your best use.


By using metadata, I’ve found that I no longer need to worry about folder hierarchies. In fact, I’ve completely changed the way that I work with groups/folders: rather than using them to organize and archive my files, I now use groups more or less exclusively as workspaces for current projects, leaving keywords and labels to do my archiving for me.

When I’m done with a project, I type two or three keywords (e.g. topic, sub-topic, client) into the comments pane of each file, then colour code the file (e.g. blue for writing, green for reference, etc). Finally, I ungroup the files and dump them in my catch-all “Archive” folder, which contains thousands of files (and no nested groups).

This way, I never need to hold a folder hierarchy in my head - and yet I know that I can always find the files that I need (without having to drill down a series of nested groups). In a sense, my original ‘workspace’ folders are never lost; all I need do is search ‘Comments’ for the relevant client and topic, then group all the retrieved files… hey presto, I have my original workspace folder back again.

This system might not work for everyone, but it’s certainly been refreshing for me. Even if you do want to create a complicated hierarchy, it’s a good idea to make use of the comments pane and the colour labels - that way it’s very hard to lose anything, even when your memory fails you.

I’ll comment specifically on your last point. Read the sections in the manual on linking and indexing. I for one fell for that early and ended up losing quite a lot of data that I couldn’t regenerate.

How do ‘power users’ use indexing or linking? I basically have my setup to copy files into the database folder so I can archive the entire db and move it around, back up at will. I’ve heard that other people just link to their file hierarchy (which sounds like twice the work to me). I need to do some more reading in these forums to get a better idea of the ‘preferred’ way…

I use a system inspired by David Allen’s Getting things done. As a rule, I store a copy of each document in a folder named “reference materials”. I then use replicants to duplicate the relevant documents which also need to be in other folders.

Some screenshots are available at

I have posted the template at the following address, if this is of interest.

I must say that I have never heard from anyone else using it so far! :open_mouth:
:unamused: :laughing:

I’ve been using your template for about two weeks. Unfortunately, I’m not well versed enough with the GTD system to feel like I can speak about it intelligently. I do appreciate the work you’ve done on the template, flow chart, and blog postings.