Hi R-helpers, I have a data.frame with three columns (lots more reps though in each), like so: 'FOO' 'BAR' 'RESULT' 1 .01 75 1 .05 12 1.1 .01 100 1.1 .05 50 1.2 .01 75 1.2 .05 12 I am calling boxplot(RESULT ~ FOO:BAR, ...) This gives me the box plots I want, but on the x-axis my labels are "1.01", "1.05", "1.1.01", "1.1.05", "1.2.01", "1.2.01". I would like to separate the factors by something other than a dot for obvious reasons. I would also like to *avoid* using the 'names' parameter to boxplot (because I am lazy and want a general solution). Please cc me directly as I read the list on digest Thanks again to such a helpful list! W
label separators in boxplots with conditioning
2 messages · wwsprague@ucdavis.edu, Gabor Grothendieck
Webb Sprague <wwsprague <at> ucdavis.edu> writes:
:
: Hi R-helpers,
:
: I have a data.frame with three columns (lots more reps though in each),
: like so:
:
: 'FOO' 'BAR' 'RESULT'
: 1 .01 75
: 1 .05 12
: 1.1 .01 100
: 1.1 .05 50
: 1.2 .01 75
: 1.2 .05 12
:
: I am calling boxplot(RESULT ~ FOO:BAR, ...) This gives me the box plots
: I want, but on the x-axis my labels are "1.01", "1.05", "1.1.01",
: "1.1.05", "1.2.01", "1.2.01". I would like to separate the factors by
: something other than a dot for obvious reasons. I would also like to
: *avoid* using the 'names' parameter to boxplot (because I am lazy and
: want a general solution).
1. Assuming BAR is numeric and always positive and the data frame is z,
this will insert a minus sign giving you some separation:
boxplot(RESULT~FOO:I(-BAR),data=z)
2. You could consider replacing BAR with something along the lines of:
z$BAR <- factor(paste("(",z$BAR,")",sep=""))