R devel - Jul 1999 - [R] R: ts - objects (PR#228)

Marcus Eger <marcus.eger@physik.uni-marburg.de> writes:
> > time(sqrt(arrts))
> Time Series:
> Start = c(1, 1) 
> End = c(5, 1) 
> Frequency = 1 
> [1] 1 2 3 4 5
Looks like a bug...
>
----------------------------------------------------------------------------
> 2. (At least) boolean indexing with matrices does not seem to work
> properly:
>
----------------------------------------------------------------------------
> > arrts[is.finite(arrts)]
> Error: subscript (10) out of bounds, should be at most 5
> > 
> > unclass(arrts)[is.finite(arrts)]
>  [1]  1  2  3  4  5  6  7  8  9 10
Yes, we discussed that a short while ago on R-devel. It's trying too
hard to retain the ts attributes in these cases.
> 
>
----------------------------------------------------------------------------
> This problem arose when I tried to plot my time series: > plot(arrts)
> It was easy to fix in plot.ts:
>
----------------------------------------------------------------------------
> ....
>   if (is.null(ylim)) 
>     ylim <- range(unclass(x)[is.finite(x)])  #by ME, 14-07-99
> #    ylim <- range(x[is.finite(x)])
>   plot.new()
> ....
...although range(x,finite=T) is probably a better fix.
> 
>
-----------------------------------------------------------------------------
> 3. Another problem had to do with the value of delta and/or start:
>
-----------------------------------------------------------------------------
> > time(ts(c(1,2,3,4),start=3,delta=1.5))
> Error: invalid start
or just: 
> ts(c(1,2,3,4),start=3,delta=1.5)
Time Series:
Start = c(3, 1) 
End = c(7, 1) 
Frequency = 0.666666666666667 
Error: invalid start

..which is pretty darn weird! CC'ed to R-bugs.

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk)             FAX: (+45) 35327907

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To:
r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

> Marcus Eger <marcus.eger@physik.uni-marburg.de> writes:
>> 
>>
-----------------------------------------------------------------------------
>> 3. Another problem had to do with the value of delta and/or start:
>>
-----------------------------------------------------------------------------
>> > time(ts(c(1,2,3,4),start=3,delta=1.5))
>> Error: invalid start
> 
> or just: 
> 
>> ts(c(1,2,3,4),start=3,delta=1.5)
> Time Series:
> Start = c(3, 1) 
> End = c(7, 1) 
> Frequency = 0.666666666666667 
> Error: invalid start
> 
> ..which is pretty darn weird! CC'ed to R-bugs.
That's just the print method for ts objects calling
time().

The underlying problem is in start().
> start(x)
[1] 3 1

I think the 2-vector form of start and end only makes sense when the
frequency is an integer. If I am right then the following versions of
start.ts and end.ts (which check that tsp[3] is an integer) should solve
the problem.

M.

"start.ts" <- function (x) 
{
    ts.eps <- .Options$ts.eps
    ##if(is.null(ts.eps)) ts.eps <- 1.e-5
    tsp <- attr(as.ts(x), "tsp")
    is <- tsp[1] * tsp[3]
    if (abs(tsp[3] - round(tsp[3])) < ts.eps && abs(is - round(is))
<
        ts.eps) {
        is <- floor(tsp[1])
        fs <- floor(tsp[3] * (tsp[1] - is) + 0.001)
        c(is, fs + 1)
    }
    else tsp[1]
}

"end.ts" <- function(x)
{
    ts.eps <- .Options$ts.eps
    ## if(is.null(ts.eps)) ts.eps <- 1.e-5
    tsp <- attr(as.ts(x), "tsp")
    is <- tsp[2] * tsp[3]
    if (abs(tsp[3] - round(tsp[3])) < ts.eps && abs(is - round(is))
<
        ts.eps) {
        is <- floor(tsp[2])
        fs <- floor(tsp[3] * (tsp[2] - is) + 0.001)
        c(is, fs + 1)
    }
    else tsp[2]
}

M.
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To:
r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._