On May 15, 2009, at 6:57 AM, deanj2k wrote:
hi everyone, wondering if you could help me with a novice problem.
I have a
data frame called subjects with a height and weight variable and
want to
calculate a bmi variable from the two. i have tried:
attach(subjects)
bmi <- (weight)/((height/100)^2)
but it comes up with the error:
Warning messages:
1: In Ops.factor(height, 100) : / not meaningful for factors
2: In Ops.factor((weight), ((height/100)^2)) :
/ not meaningful for factors
I presume that this means the vectors height and weight are not in
numeric
form (confirmed by is.numeric) so i changed the code to:
bmi <- (as.numeric(weight))/((as.numeric(height)/100)^2)
but this just comes up with a result which doesnt make sense i.e.
numbers
such as 40000 within bmi vector. Ive looked at
as.numeric(height)/as.numeric(weight) and these numbers just arnt
the same
as height/weight which is the reason for the incorrect bmi. Cant
anyone
tell me where I am going wrong? Its quiet frustrating because I cant
understand why a function claiming to convert to numeric would come
up with
such a bizarre result.
That 'height' is a factor suggests that you imported the data using
one of the read.table() family of functions and that there are non-
numeric characters in at least one of the entries in that column.
Since 'height' is a factor, if you use as.numeric(), you will get
numeric values returned that are the factor level numeric codes and
not the expected numeric values. That is why you are getting bad
values for BMI.
See:
http://cran.r-project.org/doc/FAQ/R-FAQ.html#How-do-I-convert-factors-to-numeric_003f
If you use something like:
grep("[^0-9\\.]", height, value = TRUE)
that should show you where you have non-numeric values in the 'height'
column. That is, entries for 'height' that contain characters other
than numeric or a decimal. Foe example:
height <- factor(c(seq(0, 1, 0.1), "1,10", letters[1:5]))
> height
[1] 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1,10 a
b c d e
Levels: 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1,10 a b c d e
> as.numeric(height)
[1] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
> grep("[^0-9\\.]", height, value = TRUE)
[1] "1,10" "a" "b" "c" "d" "e"
I would also check the 'weight' column for the same reasons, to be
sure that you don't have bad data there. Another approach would be to
use:
str(subjects)
which will give you a sense of the data types for each column in your
data frame. Review each column and take note of any columns that
should be numeric, but are factors.
See ?str, ?grep and ?regex for more information. You might also want
to look at ?type.convert, which is the function used by the
read.table() family of functions to determine the data types for each
column during import.
HTH,
Marc Schwartz
______________________________________________
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.
