![]() | |
![]() |
| | Thread Tools | Display Modes |
#1
| |||
| |||
|
| Returns the highest valid value in: |
#2
| |||
| |||
|
|
So here is bug/question #1: Can anyone explain why a global field shouldn't update in this situation? |
|
I'd rather they be global, makes more sense to me that way, but I will settle for what I can get. |
|
Guess what? They don't update as I enter new information. Oh, I can coax them into updating by flushing the cache and switching between browse and layout modes. But they don't update predicatably, not in the fields themselves, and not in the Data Viewer. So here is bug/question #2: Is there a reason why FileMaker should treat self-joins differently than standard ones? |
#3
| |||
| |||
|
|
Bill Marriott wrote: So here is bug/question #1: Can anyone explain why a global field shouldn't update in this situation? A global calc field only updates when a local record's field is specifically changed and that field is referenced by the global calc. The NumTransactions calc that you have is unstored and refers to related values; its change is only initiated by its appearance on the layout or its specific reference by a script. The global calc, meanwhile, is treated almost as a stored calc in that its mere appearance on screen doesn't trigger its update; only a physical change to a referenced local field can change it; the 'inferred' change of NumTransactions won't do it. I'm sorry this explanation isn't great. There's a word I've been trying to think of to describe the character of the NumTransactions field. 'Inferred' isn't exactly right. Suffice it to say that this is not desirable behavior in my mind, but FileMaker Inc does consider this to be expected behavior and not a bug. While I was looking forward to having global calcs in FM7, I have found them for the most part to be useless. I'd rather they be global, makes more sense to me that way, but I will settle for what I can get. Agreed Guess what? They don't update as I enter new information. Oh, I can coax them into updating by flushing the cache and switching between browse and layout modes. But they don't update predicatably, not in the fields themselves, and not in the Data Viewer. So here is bug/question #2: Is there a reason why FileMaker should treat self-joins differently than standard ones? There probably is a reason, but not a good one. I hope they fix it some day, but I am accepting that this might just be one of those things we have to live with in FM, that we will forever have to work around. If you were to somehow refer to a local field in your calc, then the calc would refresh itself any time that local field was changed. For example, make EarliestTransaction2 (calc, unstored, date result) = Min(EntriesClone::TransactionDate) + (case(TransactionDate, 0) This should force the min() calc to update each time the TransactionDate field is specifically updated in the local record. Not a perfect solution for every situation, but it might help you. I'd also report it as a problem to FileMaker http://filemaker.com/company/product/problems.html> because it really does deserve to be "fixed". ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Howard Schlossberg (818) 883-2846 FM Pro Solutions Los Angeles, California FileMaker 7 Certified Developer Associate Member, FileMaker Solutions Alliance |
![]() |
| Thread Tools | |
| Display Modes | |
| |