3 Replies Latest reply on Jan 6, 2013 5:31 PM by RickWhitelaw

    Question about database design

    MarcWhite

      Title

      Question about database design

      Post

           Hi all,

           I am an 'advanced beginner' FM Pro user who designs databases mainly for his own research. I have a tissue database that includes a lot of information about patients, tissue samples, how they are processed and stored, where they are shipped to. I separated many of the informations into their own tables so that the table number is quite large (compared to the examples I saw), and the relationship graph developed into a spider web (see picture). However, this works better for me than creating dedicated TO groups. The attached ERG for example shows everything related to patients within the red solid box, everything related to tissue samples in the black solid box, and everything related to processed samples (isolated DNA from tissue e.g.) in the red dotted box. I then added a relationship to show with whom tissue samples were shared. A dedicated TO group for that is shown in the lower right of the ERG, and I can work with that. But if I wanted to create a report for a collaborator to show for example which liver samples came from patients of a specific race and were treated in a certain way, I would have to add most of the TOs that are already present in the large TO group. I often come up with new questions to look at the information, and adding TOs to the big ERG works fine (in the latter case - shipping report - TOs in the black dotted box).

           However, as I would like to improve my database/FM skills I am interested in what you guys think or would suggest. The examples I looked into and the examples from "FM Pro - The Missing Manual" and "The Filemaker Bible" have fewer tables (use value lists instead of separate tables), and seem to advice against such an approach.

           Kind regards,
           Marc

      Graph.png

        • 1. Re: Question about database design
          philmodjunk

               Compared to some of my systems, this is a very small number of table occurrences.

               What you have seems quite close to Anchor Buoy. Relatively few changes would be needed to implement that method for organizing relaitonships--should you choose to do so. In a sense, each boxed set of table occurrences could be set up as a group with just a few additional TO's outside of the box included.

               Please note that I really can't make a detailed study of your Relationships window as I can't read the text even when I attempt to zoom in to see the details.

          • 2. Re: Question about database design
            MarcWhite

                 Thank you for the kind reply. That sounds encouraging. The picture I posted was indeed of low resolution. I attached a version with better resolution. I would certainly appreciate some more input if possible. Thank you.

                  

            • 3. Re: Question about database design
              RickWhitelaw

                   The relationship graphs from the multi file solution that I designed to run my business look like plates of spaghetti compared to your well organized graph, and that's after organizing it! Nothing to worry about. My solution runs fine.