2 Replies Latest reply on Sep 27, 2012 4:39 PM by cortical

    How to choose to add new table vs new database (related dbs)?

    snickwit

      Hello,

       

      I haven't found anything that discusses when one might decide to add related data as a new table in existing database or new database. My general thinking would be to create a new table if it's related data but how does one determine or make the asessment if that might make the database too bloated or affect performance? What are the cons or impact of adding as a new table vs new database? For example, what's impact if you decide to add as table in existing database but later have seperate database and now need to relate to the table in other database?

        • 1. Re: How to choose to add new table vs new database (related dbs)?
          gdurniak

          In the old days, this was never an issue:  one file, one table.  Simple.  Now you have choices

           

          There are many threads here, that discuss the pros, and cons

           

          I have never seen a performance difference. It's all about convenience. It's just easier to develop with everything in one place

           

          if you have one table you know will grow huge, it might pay to make that a separate file ( and you can change your mind later )

           

          greg

           

          > how does one determine or make the asessment if that might make the database too bloated or affect performance?

          • 2. Re: How to choose to add new table vs new database (related dbs)?
            cortical

            Sounds like you need to be searching for 'data separation model'

             

            Myself, I put 'all' tables in a single b_data file , exept for docs/images in a separate file, and occasionally LUT data in a dedicated file when there is periodic updates e.g goverment PBS benefits; so I can import the latest yearly/whenever schedule into a copy, and email the update so the user can drag and drop replace/upload...