May 19, 200619 yr Hi, I recently converted a databease created in FM6 to FM8. Its new, so only has a handful of records (sample data only). I converted it to FM8 and it seems to run slower. I can see the records being "painted" on the sceen. My question is whether this is because this database was converted from FM6, or would it be the same if the database was created from scratch using FM8? tks
May 19, 200619 yr would it be the same if the database was created from scratch using FM8? Probably, because what pulling the teeth out of your solution, is a struggle to make sense of the file references. Other issues exists though, if you solution is somewhat relation shy, will it probably contain a lot of chained unstored calcfields as well as being splayed with dozens of Summary fields shown elsewhere than dedicated summary reports. Unfortunately is the reason "paining" that you probably tries to beat filemaker out of the database realm into being a spreadsheet with an interface layer, and the term normalization is what's required. --sd
May 19, 200619 yr Author Soren, I'm not quite sure what you mean about me trying to "beat filemaker out of the database realm into being a spreadsheet with an interface layer".. My system is an order management and inventory system with about 20 files. There are a lot of calculations in the inventory file for caculating inventory balances... but even just displaying a list of orders, FM8 is noticeably slower.
May 19, 200619 yr There are a lot of calculations in the inventory file for caculating inventory balances Exactly, databases are for storing data ...you need to take a closer look (... or here listen) at topics like "transaction model" http://podcast.Adatasol.com/media/Filemaker_PodCast_03262006.mp3 Or read on here: http://en.wikipedia.org/wiki/Database_transaction Out of curriosity would I like to know what you think the script commands: Commit Record/Request Open Record/Request Revert Record/Request ...as well as the calc'function: GetNextSerialValue ( fileName ; fieldName ) ...if it weren't to support a more ACID compliant model? There have been attampt in this direction with earlier versions: http://www.filemakerpros.com/LULAST.zip Which bring up an extra clue, for your debugging process - today should the lookup calc'function be utilized in autoenter fields instead of genuine lookups. --sd
May 19, 200619 yr Author Soren, I get the impression you are slapping me but its hard to tell as I don't fully understand why you would ask me what a commit record request is for. My database stores transactions and also makes calculations based on those transactions. If you have an order database with transactions (incoming and outgoing stock) how would you determine the current balance of an item without performing calculations? I realize that all these calculations will slow my system down, if there is a way to avoid using so many calculations, I'd love to hear it. thanks david
May 31, 200619 yr Hi David, Do you still have the solution in multiple files or have you consolidated the tables into one file? Calcs should only trigger when called to be displayed, printed or used in a script or other calculation. You might check the file references to see if there are problems resolving the file locations. (file/define/file references) Also, you might consider the number of fields filemaker is trying to index. This can cause performance issues, because filemaker must build the index for each indexed field when a file is opened. Did you convert the whole solution together or one file at a time? Tim Edited May 31, 200619 yr by Guest
Create an account or sign in to comment