I follow this forum quite closely and see that many of you use Brilliant Database for a variety of applications. Data-logging appears to be straightforward and not too problematic but I am curious about those of you that use Brilliant Database for complicated analyses and reporting. More specifically, I am interested if you go back to the raw data and verify the results manually or through another program such as EXCEL. I have always verified my information in this fashion just to ensure that the results I am seeing through Brilliant Database are correct. This causes me problems!
What troubles me is when Brilliant Database gives different results when the data has NOT changed at all. The IT Department at the Hospital contacted me Friday with a new request. I pondered over how to achieve this new information over the weekend and adjusted the program to give the desired analysis. No scripting changes were required nor were there any modifications to the Hospital data - I had set up the project so that the user can create new parameters dynamically. The first time I attempted to run this, it gave a null solution - which was incorrect. I looked over the changes and verified that EVERYTHING was correct and ran it again. The second attempt produced the proper results (this was verified by downloading all the data to EXCEL and checking). I ran it a third time and the results were identical and correct; however, running it a fourth time resulted in a null solution (it showed 0's). There were absolutely no changes made to any of the data as I have not even looked at this program in about three months.
This morning, I have run the application five times and three times the solution was perfect whereas the other two resulted in 0's... and no changes were ever made to the data or scripting??? I have seen this happen before with other things I was working on and I am really, really curious if others have experienced results that change when no changes were made. Do others always verify the accuracy of the results of Brilliant Database?
It would be very easy to dismiss this problem as one caused by me by poor scripting practices but I feel very confident this this is not the issue. Maybe I am trying to use Brilliant Database for applications that it cannot handle but I am doing nothing more than using queries (which I have demonstrated repeatably no not always work properly), simple comparisons (does a=b), and simple addition (c=a+b). I keep trying to use this program but my confidence is waning...
I've worked on so many projects of all different sizes that I've seen some odd things happening for sure.
These days though when ever I see funky things happened I can nearly always put it down to a query I've written is finding too many records that match the search. Or the user input finds too many records. It's always “too much” of something found.
To solve this, I either rewrite the query so it's more accurate (finds less results that better match), or if there is always the possibility of the end user entering in something that will product a high number of hits, I'd break down a Query into a number of them.
For example, I have one database that needs to read back though 80,000+ records often. In this case rather than running one query, I've copied and pasted the same query 26 times and I use an IF statement to look at the first letter in the alphabet. Basically all I'm doing is spreading out the load as a query seems to have this magical upper unknown limit where things get strange.
If Brilliant Database were responsive to emails, I'm pretty sure I can list out a nice few “funky” things this software does when it's under heavy load that could easily be solved, rather than having to code you way around them.
It's awesome software, it just needs the developer Mikhail to put in some effort to address certain issues.