Re: Database design for scientific tests

[Al Lombardo]

Usually if you have a situation where you'll need 0.1 or more instances of data for each row, you should break it out into another table. So in essence, you would set up a table of all possible initial conditions (as a look up table) where pressure, temp etc. would be listed, then in another table (say initial_test_conditions) have a link between test and initial condition and add a value field for the condition. Initial_test_conditions would hold only those conditions that are important for that instance, not all possible combinations between initial conditions and test.