[SciPy-User] failing to create a scikit.timeseries object

Pierre GM pgmdevlist@gmail....
Wed Jun 2 14:50:26 CDT 2010


On Jun 2, 2010, at 3:03 PM, Robert Ferrell wrote:

>  I think that second resolution is the highest frequency available (at least that's what I see in the documentation).

Yes. 


> Here's how to make a date array:
> 
> ts.date_array(freq='S', dlist=d)
> DateArray([01-Jan-2010 10:10:02, 01-Jan-2010 10:10:02],
>           freq='S')
> 
> I don't know if timeseries Date objects can resolve microsecond differences.  

It shouldn't. Internally, dates are stored as integers. The difference between one frequency and another is the reference used to convert integers back to some dates (eg., with an "ANN" frequency, we just store the year; with a "MON" freq, we just store the nb of months since 01/01/01...). Because so far the highest frequency is "SEC", there's no way to distinguish times at a larger frequency.


More information about the SciPy-User mailing list