![]() | |
![]() |
| | Thread Tools | Display Modes |
#1
| |||
| |||
|
#2
| |||
| |||
|
|
I'm having a problem with autoentered serial numbers. The database is multi user, hosted on FMS7. Users create new contact records by clicking a button, which runs a script to create a new records in the contacts table. Out of ~550 records, I've found two pairs of records with duplicate serial numbers. The users can't edit the serial numbers; it looks like FileMaker is assigning them the same serial number if they're created at the same time. Has anybody seen this with 7 before? |
|
For now, I've put in a random delay before creating a record... |
#3
| |||
| |||
|
|
In article <020920050832393304%BrianHall (AT) nospam (DOT) pobox.com>, BrianHall (AT) nospam (DOT) pobox.com says... I'm having a problem with autoentered serial numbers. The database is multi user, hosted on FMS7. Users create new contact records by clicking a button, which runs a script to create a new records in the contacts table. Out of ~550 records, I've found two pairs of records with duplicate serial numbers. The users can't edit the serial numbers; it looks like FileMaker is assigning them the same serial number if they're created at the same time. Has anybody seen this with 7 before? I've never seen it or even heard of it EVER in all the time I worked with ANY version of Filemaker. I've even done concurrency tests. (Setting up multiple PCs on a looping script that just created piles of records). And its never happened. (And I let it run 10s of thousands of record.) (I was testing something else, but it relied on unique serial numbers.) Are you *sure* there is no other possible cause. If it really is happening than I'd suspect some sort of database corruption, not a defect in the filemaker software itself. But I'd be much more inclined to think that its buggy script that's applying a setfield where it shouldn't or something. Or perhaps goofups during an import? For now, I've put in a random delay before creating a record... That shouldn't make any difference. The odds of two users pushing 'new' at the same time are identical to the odds of two users pushing 'new' at a ramdomized interval. Either way sooner or later the new record commands will arrive at the server 'at the same time'. Except that there is no such thing as 'as the same time'. Filemaker sequences new record requests. One must happen before the next. A couple questions/suggestions: 1) Are the 'duped' records adjacent in the database? 2) Add a timestamp to the table so that you can see what time they were created if it happens again. 3) Try to force it to reproduce it. (mulitple computers looping the new record script...) If you can make it happen that will help immensely to diagnose it. 4) Add a 2nd independant autoenter serial. See if future 'duped' records are duping both. -regards, Dave |
![]() |
| Thread Tools | |
| Display Modes | |
| |