Thanks for the reply Peter.
On 18 January 2011 22:52, Peter Ehlers<ehlers at ucalgary.ca> wrote:
Since you don't provide data, let's borrow from the
help(droplevels) page:
I had no joy with my R install finding droplevels exactly, but found this
instead:
gdata::drop.levels Drop unused factor levels
Is that the same?
aq<- transform(airquality,
Month = factor(Month, labels = month.abb[5:9]))
str(aq)
#'data.frame': 153 obs. of 6 variables: |
# $ Ozone : int 41 36 12 18 NA 28 23 19 |
# $ Solar.R: int 190 118 149 313 NA NA 29 |
# $ Wind : num 7.4 8 12.6 11.5 14.3 14. | etc
# $ Temp : int 67 72 74 62 56 66 65 59 |
# $ Month : Factor w/ 5 levels "May","Jun |
# $ Day : int 1 2 3 4 5 6 7 8 9 10 ... |
Now see if the following give you some R inspiration:
plot(Ozone ~ Temp, data = aq)
This highlights one of the very confusing aspects of R language for me; is
plot(x, y) the same as plot (y ~ x)? Seems to be, but maybe I'm missing some
nuance.
plot(Ozone ~ Temp, data = aq, subset = {Month == "Sep"})
boxplot(Ozone ~ Month, data = aq)
boxplot(Ozone ~ Month, data = aq,
subset = {Month != "Aug"})
boxplot(Ozone ~ Month, data = aq,
subset = {!(Month %in% c("Jul", "Aug"))})
boxplot(Ozone ~ Month,
data = droplevels(subset(aq, subset = {Month != "Aug"})))
boxplot(Ozone ~ Month,
data = droplevels(subset(aq, !(Month %in% c("Jul", "Aug")))))
Thanks for these examples, they mostly make sense!
BTW, attach() is not usually a good idea; have a look at ?with.
Great, I thought I had that trick nailed. Obviously there needs to be an R
equivalent of l2tabu.
Cheers,
Ben.
[[alternative HTML version deleted]]