Internaly stored record numbers missbehaving

[charper]

I'm having some terrible problems with filemaker pro 4.1's internally stored record numbers.

My understanding of this special record number field is that Filemaker is supposed to generate a new unique value for it for every record that gets created in the database. I also expect it to increment serially through the entire life of the database. But instead what I see is that when a new record gets created, sometimes the record number for other records in the database get "pushed up" by +1 and the new record is assigned a previously used record number. Example:

John Smith=1

Bob Jones=2

Lisa Clay=3

Tim Riley=4

Create a new record: Sharon Taylor

You get:

John Smith=1

Bob Jones=2

Sharon Taylor=3

Lisa Clay=4

Tim Riley=5

The insertion point for the record seems to be arbitrary. What logic is filemaker using here?

Also, eventually my records end up sharing the same record numbers:

Sharon Taylor=3

Harold Johnson=3

Barbara Bonehead=3

How is this possible? Is this a Filemaker 4.1 bug? What's the point of having non unique record numbers?

Last but not least, sometimes when I export/import records from this database (database A) into another database (database : the record numbers in database B jump to astronomically high levels. For example:

In database B the last record number generated was 387. I import a few record into it from database A (using intermediate Tab files). After the import, the imported records have record numbers like 12,982,332!!

What in the hell is Filemaker doing?

This is about ready to kill us because database B is being web published via lasso and this system requires that this record number remain tame (i.e. behaves more or less like a regular auto-enter numeric serial number). Please help.

[BobWeaver]

Make sure you're not confusing record number with record ID. The record number is the position in the current found set. The record ID is a unique number. The record number can and will change depending on how you sort and your found set.

[This message has been edited by BobWeaver (edited March 01, 2001).]

[charper]

My god, your right. I'm looking at record number and thinking that I'm looking at record id. At least that's my error as far as database A is concerned.

Ok. Now I'm looking at RECORD ID in

database A and it looks perfectly fine - behaves as expected. BUT.....

its still the case that after record are exported/imported to database B, RECORD ID's in database B shoot through the roof i.e. 25,109,933 (when they were down around 40,000 before the import). Now, as I said before, we're also interfacing database B with a lasso driven website. Is it possible that our web developer's Lasso code is doing something crazy that would cause RECORD IDs to sky rocket in this manner? Among other things, this lasso code IS creating new records in database B. After I do an import from database A, however, our website developer says that's when Lasso starts pulling back outrageously high RECORD IDs from database B.

I realize you may not be a Lasso expert, but any thoughts on this would be greatly appreciated. We'll be visiting some Lasso forums and asking about this over there. Thanks very much for your help.

[BobWeaver]

I'm kind of vague on this. I know I read somewhere that the record ID will not necessarily run in sequence depending on how the record is created, but I can't remember what causes jumps in the numbering. It could be that creation of records from related files causes this. In any event, you are not supposed to rely on the record ID's being in any particular pattern. You may have to invent your own unique numbering scheme if you don't like how the record ID's are assigned.