Gads, don't you just love inheriting applications/data with no descriptors?

If and only if those 9 characters are the same in every records, then I'd be tempted to #define it as a constant, just plug it into new records whilst I madly ran around trying to find someone who might have a clue on it. Not knowing what kind of workfield (accounting, health, etc) you're dealing with, I wouldn't feel safe to make a WAG.

>Thanks Dorris
>
>any idea what the first part could be ?
>
>>Then I'd bet that the latter part of that is day(date()) + seconds() * 10000 + '0000'
>> Or, some minor variation thereof
>>
>>>pretty sure its a timestamp. The 231258 line was when I set my clock date to 23rd Feb
>>>
>>>T1712261622122531240000
>>>T1712261622124027560000
>>>T1712261622124938660000
>>>T1712261622125659720000
>>>T1712261623125813170000
>>>T1712261622010219640000
>>>
>>>>It seems to me that the 2212253etc is a little long for a time stamp? Obviously it's right padded with zeroes, but wouldn't a time stamp look more like 221253 (hour,minute,seconds) or 81517894 (seconds() with the decimal removed). I'd have to see a few more record keys in the order they appear to make more sense of it.
>>>>
>>>>I know - not much help
>>>>
>>>>>This is a a long shot.
>>>>>
>>>>>I'm dealing with a package and a table in that package has keys generated like.
>>>>>
>>>>>T1712261622122531240000
>>>>>
>>>>>I think that 22122531240000 is some sort of time stamp but I'm not sure about the T17122616. Its the same for all the entries I have put in. Could it be some sort of connection number ? It doesn't alter if I log out of the system and log back in.
>>>>>
>>>>>I want to be able to generate some data to enter into this table and I'd like to emulate the existing key structure.