On Jan 24, 2012, at 4:25 PM, Curt Seeliger wrote:
Yes, you're right about this being a floating point issue. I guess I
wasn't clear enough that this was already understood. I'd have
responded
earlier the response somehow missed my mailbox.
My question is rather whether there is a work aro
Yes, you're right about this being a floating point issue. I guess I
wasn't clear enough that this was already understood. I'd have responded
earlier the response somehow missed my mailbox.
My question is rather whether there is a work around for correctly
displaying POSIXct values as charact
This is basically FAQ 7.31. WIth floating point number, you have
about 15 digits of significance, so if you look at the value:
>> as.numeric(as.POSIXct('2010-06-03 9:03:58.324'))
> [1] 1275581038.3239998817
when you get down to the milliseconds, this is about as much accuracy
as you will get bas
First, the reproducable example, showing how converting from character to
POSIXct to character changes the milliseconds in the first time stamp
though not in the second:
> as.POSIXct('2010-06-03 9:03:58.324')
[1] "2010-06-03 09:03:58.323 PDT"
> as.POSIXct('2010-06-03 9:03:58.325')
[1] "2010-06-
4 matches
Mail list logo
