Mysterious rounding of seconds problem.

Discussion created by jacobne on Dec 6, 2012
Latest reply on Dec 7, 2012 by gis_tech
Hello All,

I'm experiencing a strange "rounding of seconds" problem during some date conversions and was curious to know if anyone had any theories as to what may be happening.  Any suggestions would much appreciated.

I've been converting timestamps from a csv file into an ArcGIS 10 geodatabase feature class and have found that minor inconsistencies in the seconds are occurring.  Some records have timestamps that are off by +1 second.

CSV File                         
2012-12-05 14:30:10 UTC
12/5/2012 2:30:11 PM

At first I suspected microseconds, but I've checked and rechecked, and there are definitely no microseconds.  What's really odd is if I view the table in the GDB it looks correct, however if I print the values with a search cursor in the Python Window I see minor errors in the seconds.  How is something like that possible?  Is this a problem with the search cursor or the table in the GDB?