[R] ts format for daily time serie

[Achim Zeileis]

On Thursday 20 November 2003 19:53, Kenneth Cabrera wrote:

> Prof Brian Ripley <ripley at stats.ox.ac.uk> wrote:
> > It's probably best to handle this an an x-y plot (?plot.POSIXct)
> > or as an irregular time series (see package its on CRAN).
>
> A daily time serie is not a regulary spaced time serie?
>
> > The ts function is designed for time series that you want to do
> > time-series analysis on.
>
> That is what I want, make time-series analysis, but at the same
> time I want to plot it directly.
>
> Achim Zeileis <zeileis at ci.tuwien.ac.at> wrote:
> > ts() only allows regulary spaced time series. Try the package its
> > or the irts() function in package tseries for irregularly spaced
> > time series.
> >
> > hth,
> > Z
>
> A daily time serie is not a regulary spaced time serie?

Not if it has gaps like in the excerpt of your data given below. 
Furthermore the frequency is not exactly 365.
Z

> >> Excerp of the database:
> >>
> >>            FECHA     TRM
> >> 1    01/01/2000 1873.77
> >> 2    02/01/2000 1873.77
> >> 3    03/01/2000 1873.77
> >> 4    04/01/2000 1874.35
> >> 5    05/01/2000 1895.97
> >> .
> >> .
> >> .
> >> 1397 10/11/2003 2843.82
> >> 1398 11/11/2003 2840.41
> >> 1399 12/11/2003 2840.41
> >> 1400 13/11/2003 2845.69
> >> 1401 14/11/2003 2850.24
> >> 1402 18/11/2003 2842.53
> >> 1403 19/11/2003 2831.97
> >> 1404 20/11/2003 2826.60
>
> Using M2, Opera's revolutionary e-mail client:
> http://www.opera.com/m2/