Bug introduced in 10.1 or earlier and fixed in 10.2
DateObject is new in 10.0
DateObject was introduced in version 10 of Mathematica. One can call it to fetch the current local date and time:
By default, Mathematica uses your local time zone. One can also, according to the documentation, specify a time zone via (to give the documentation example):
DateObject[TimeZone -> 0] (* Fri 17 Apr 2015 22:11:20 GMT *)
Here, we get the current date and time in GMT. So, at the time I write this, it is 2 hours to midnight on Friday 17 April in Greenwich. But what time is it in Tokyo? Well, everybody knows that Tokyo is 9 hours later than Greenwich, so it should be 7 in the morning on Saturday 18 April, yes? Mathematica disagrees:
DateObject[TimeZone -> 9] (* Sun 19 Apr 2015 07:20:20 GMT+9 *)
Sunday! I get the right time, but the wrong date. This is odd! I can check the other direction, too: 24 hours ago, it was 2 hours to midnight on Thursday 16 April in Greenwich, so I run
DateObject[TimeZone -> -24] (* Wed 15 Apr 2015 22:25:40 GMT *)
and find out that it was, in fact, Wednesday!
So, it appears, that if one specifies a
TimeZone for the current date that belongs to the next or previous day,
DateObject skips over a whole day when returning the answer. I've checked that this behavior repeats each time it crosses through midnight.
If one explicitly passes
DateObject there are no problems:
DateObject[Now, TimeZone -> 9] DateObject[Now, TimeZone -> -24] (* Sat 18 Apr 2015 07:28:46 GMT+9 *) (* Thu 16 Apr 2015 22:28:46 GMT *)
DateObject just fine:
TimeZoneConvert[DateObject, 9] TimeZoneConvert[DateObject, -24] (* Sat 18 Apr 2015 07:31:38 GMT+9 *) (* Thu 16 Apr 2015 22:31:38 GMT *)
Any ideas what is going on? I'm working in version 10.1.0.0 on Mac OS X. According to the comments, it can be reproduced on other systems. Am I just abusing the
TimeZone syntax? Since the date and time functionality is so new, I haven't been able to find any similar problems posted yet.