[R] unexpected result from format(x, digits)

[(Ted Harding)]

On 28-Apr-10 21:45:12, Steve Taylor wrote:
>  
> Is this a bug somewhere?  The format function, using a specific
> number of digits, doesn't give sensible results:
>  
> R> set.seed(2);print(x<-rexp(5))
> [1] .
> R> format(x,digits=1)
> [1] "1.87" "0.40" "0.15" "1.73" "0.09"
> R> format(x,digits=2)
> [1] "1.87" "0.40" "0.15" "1.73" "0.09"
> R> format(x,digits=3)
> [1] "1.8654" "0.4047" "0.1467" "1.7307" "0.0895"

Not a bug (unless you consider the documented behaviour to amount
to a bug)!

From: ?format

  digits: how many significant digits are to be used for
    numeric and complex 'x'. The default, 'NUL'?, uses
    'getOption(digits)'. This is a suggestion: enough
    decimal places will be used so that the smallest
    (in magnitude) number has this many significant digits, 

And that is exactly what is happening.

In your first example, the smallest number 0.08953 rounds
(to 1 significant digit) to 0.09 and therefore prints as 0.09.

In the second, 0.08953 rounds (to 2 significant digits) to 0.090,
which can be printed without further loss of precision as 0.09
while giving the other numbers 2 significant digits.

In the third, 0.08953 rounds (to 3 significant digits) to 0895
The point about the four significant digits given for the
other (larger numbers) is that withg format() *all the numbers
are printed to the same number of decimal places*. Hence the
smallest number gets 3 signficicant digits, hence 4 decimal
places. The other larger numbers will then also get 4.

It looks as though, lurking behind your question, is a secret
wish for a fixed number of *decimal places* (1, or 2 or 3 in
your examples). For this, try formatC() (but be careful):

  X <- c(1.86535, 0.40475, 0.14665, 1.73071, 0.08953)
  formatC(X,2)
  # [1] "1.9"  "0.4"  "0.15" "1.7"  "0.09"
  formatC(X,2,format="f")
  # [1] "1.87" "0.40" "0.15" "1.73" "0.09"
  formatC(X,3)
  # [1] "1.87"   "0.405"  "0.147"  "1.73"   "0.0895"
  formatC(X,3,format="f")
  # [1] "1.865" "0.405" "0.147" "1.731" "0.090"

Hoping this helps!
Ted.

--------------------------------------------------------------------
E-Mail: (Ted Harding) <Ted.Harding at manchester.ac.uk>
Fax-to-email: +44 (0)870 094 0861
Date: 28-Apr-10                                       Time: 23:12:53
------------------------------ XFMail ------------------------------