xscale.components
and yscale.components
arguments can be
functions that determine tick mark locations and labels given a
packet. For more direct control, the axis
argument can be a
function that actually draws the axes. The functions documented here
are the defaults for these arguments. They can additonally be used as
components of user written replacements.xscale.components.default(lim,
packet.number = 0,
packet.list = NULL,
top = TRUE,
...)
yscale.components.default(lim,
packet.number = 0,
packet.list = NULL,
right = TRUE,
...)
axis.default(side = c("top", "bottom", "left", "right"),
scales, components, as.table,
labels = c("default", "yes", "no"),
ticks = c("default", "yes", "no"),
...)
print.trellis
) is being
processed. In cases where all panels have the same limits, this
function is called only once (rattop
and right
components of the result, as appropriate. See below for
interpretation.scales
argument
supplied to the high level function, suitably standardized.xscale.components.default
and
yscale.components.default
.as.table
argument in the high level
function.scales
are used.scales
are used.xscale.components.default
and yscale.components.default
return a list of the form suitable as the components
argument
of axis.default
. Valid components in the return value of
xscale.components.default
are:num.limit
bottom
ticks
and
labels
. ticks
must be a list with components
at
and tck
which give the location and lengths of tick
marks. tck
can be a vector, and will be recycled to be as
long as at
. labels
must be a list with components
at
, labels
, and check.overlap
. at
and
labels
give the location and labels of the tick labels; this
is usually the same as the location of the ticks, but is not
required to be so. check.overlap
is a logical flag
indicating whether overlapping of labels should be avoided by
omitting some of the labels while rendering.top
TRUE
,
top
is treated as being the same as bottom
; if
FALSE
, axis annotation for the top axis is omitted.
Alternatively, top
can be a list like bottom
.yscale.components.default
are left
and right
.
Their interpretations are analogous to (respectively) the
bottom
and top
components described above.Lattice
,
xyplot
,
print.trellis
str(xscale.components.default(c(0, 1)))
set.seed(36872)
rln <- rlnorm(100)
densityplot(rln,
scales = list(x = list(log = 2), alternating = 3),
xlab = "Simulated lognormal variates",
xscale.components = function(...) {
ans <- xscale.components.default(...)
ans$top <- ans$bottom
ans$bottom$labels$labels <- parse(text = ans$bottom$labels$labels)
ans$top$labels$labels <-
if (require(MASS))
fractions(2^(ans$top$labels$at))
else
2^(ans$top$labels$at)
ans
})
## Direct use of axis to show two temperature scales (Celcius and
## Fahrenheit). This does not work for multi-row plots, and doesn't
## do automatic allocation of space
F2C <- function(f) 5 * (f - 32) / 9
C2F <- function(c) 32 + 9 * c / 5
axis.CF <-
function(side, ...)
{
ylim <- current.panel.limits()$ylim
switch(side,
left = {
prettyF <- pretty(ylim)
labF <- parse(text = sprintf("%s ~ degree * F", prettyF))
panel.axis(side = side, outside = TRUE,
at = prettyF, labels = labF)
},
right = {
prettyC <- pretty(F2C(ylim))
labC <- parse(text = sprintf("%s ~ degree * C", prettyC))
panel.axis(side = side, outside = TRUE,
at = C2F(prettyC), labels = labC)
},
axis.default(side = side, ...))
}
xyplot(nhtemp ~ time(nhtemp), aspect = "xy", type = "o",
scales = list(y = list(alternating = 3)),
axis = axis.CF, xlab = "Year", ylab = "Temperature",
main = "Yearly temperature in New Haven, CT")
## version using yscale.components
yscale.components.CF <-
function(...)
{
ans <- yscale.components.default(...)
ans$right <- ans$left
ans$left$labels$labels <-
parse(text = sprintf("%s ~ degree * F", ans$left$labels$at))
prettyC <- pretty(F2C(ans$num.limit))
ans$right$ticks$at <- C2F(prettyC)
ans$right$labels$at <- C2F(prettyC)
ans$right$labels$labels <-
parse(text = sprintf("%s ~ degree * C", prettyC))
ans
}
xyplot(nhtemp ~ time(nhtemp), aspect = "xy", type = "o",
scales = list(y = list(alternating = 3)),
yscale.components = yscale.components.CF,
xlab = "Year", ylab = "Temperature",
main = "Yearly temperature in New Haven, CT")
Run the code above in your browser using DataLab