[R] Bug? in summary( ) function base package

[hadley wickham]

On Wed, Apr 30, 2008 at 1:06 PM, Richard Saba <sabaric at auburn.edu> wrote:
> There seems to be an error in the summary() function when applied to "ts"
>  class objects. The results of a call to summary( ), on the R  "ts" data set
>  USAccDeaths , reports the wrong value for Max. The value reported by the
>  summary function is 11320. The max( ) function returns the correct value
>  11317, the July 1993 value. Coercing the data to a data.frame and calling
>  summary returns the correct max value. A search of R -help found a post in
>  2007 that mentioned a problem but attributed it to rounding errors. But this
>  is too large a difference to account for a simple rounding error.

Notice that all the numbers in :

>  > summary(USAccDeaths)
>    Min. 1st Qu.  Median    Mean 3rd Qu.    Max.
>    6892    8089    8728    8789    9323   11320

have four significant digits and then carefully read the documentation
for the digits argument of summary.  That should suggest both the
cause and a solution.

Hadley


-- 
http://had.co.nz/