Thank you all for helping me out.
As Michael points out, I abused the rounding and formatting of print()
while debugging.
The default number of digits to print is 7 according to ?print.default,
which makes floating point numbers to be somewhat plausible for index vector
subsetting at first glance:
------------------------------------------
> print(0.000000000001)
[1] 1e-12
> print(0.999999999999)
[1] 1
------------------------------------------
However, since ?[ mentioned that numerics are coerced to integers in fact, it
turns out to be an floating point issue.
As Berend and Rui noted, it can be revealed by:
------------------------------------------
> print(0.000000000001, digits = 17)
[1] 9.9999999999999998e-13
> print(formatC(0.000000000001, format="f", 17))
[1] "0.00000000000100000"
> print(0.999999999999, digits = 17)
[1] 0.99999999999900002
> print(formatC(0.999999999999, format="f", 17))
[1] "0.99999999999900002"
------------------------------------------
As we can see above, floating point number is a miracle, and there is still
some magic with print(). :)
BTW, I think this is a general issue, which should be carefully considered
regardless of R or other languages.
Have a nice day.
Cheers,
Guo
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
