Archives of the TeradataForum
Message Posted: Fri, 13 Jun 2008 @ 13:47:14 GMT
Dieter has already confirmed that we are not missing out on the storage side of timestamps.
Where we seem to be missing out with Teradata compared to other RDBMS is the precision of time interval measurements (cases when we use the database as time measurement device, not just a place we store timestamps collected by other systems). When I say "missing out", it does not necessarily mean a big loss but you know how it is these days, you have to do certain things just because you can do them or because others do them.
The competition is ahead. Oracle routinely provides 6 decimal digits on Unix and 3 on Windows servers. It does not mean that all these measurements are really precise but it gives a feeling that you get more for less.
Teradata is a special case in this department:
It has to deal with many system clocks (one per node). Perhaps this is the underlying reason why they don't want to be precise. Clock synchronization between servers up to a reasonable level is cheap but synchronization up to the 6th decimal place, if at all possible, can probably demand a price with 6 digits *before* the decimal separator. So, the reasoning can be: if we can't synchronize system clocks then why bother providing irrelevant decimals?
This is just my guess. Teradata insiders may provide totally different views. As always, it would be interesting to hear their comments on the current situation and the future plans.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|