At 04:23 PM 21/10/2010, Kjell Rilbe wrote:

>Helen Borrie skriver:
>> The difference between two timestamps returns a result in days. 0.0002
>> of a day is 17.28 seconds....but the interval calculation corrects to 4
>> places of decimal. So, for example, if the elapsed time was actually
>> 16.703 seconds, the corrected result in days would be still be 0.0002.
>
>Oh. But does datediff provide better precision?

Maybe...or not...probably the only one who really knows about the precision of the function's output is Adriano, who wrote the code. Did you test?

Did you happen to notice the OP was asking about why his timestamp math wasn't giving him the result he expected, not about how to use datetime functions?

./heLen