Default Trellis Strip Function

D_strip.default {lattice}

R Documentation

Default Trellis Strip Function

Description

strip.default is the function that draws the strips by default in Trellis plots. Users can write their own strip functions, but most commonly this involves calling strip.default with a slightly different arguments. strip.custom provides a convenient way to obtain new strip functions that differ from strip.default only in the default values of certain arguments.

Usage

strip.default(which.given,
              which.panel,
              var.name,
              factor.levels,
              shingle.intervals,
              strip.names = c(FALSE, TRUE),
              strip.levels = c(TRUE, FALSE),
              sep = " : ",
              style = 1,
              horizontal = TRUE,
              bg = trellis.par.get("strip.background")$col[which.given],
              fg = trellis.par.get("strip.shingle")$col[which.given],
              par.strip.text = trellis.par.get("add.text"))
strip.custom(...)

Arguments

`which.given`	integer index specifying which of the conditioning variables this strip corresponds to.
`which.panel`	vector of integers as long as the number of conditioning variables. The contents are indices specifying the current levels of each of the conditioning variables (thus, this would be unique for each distinct packet). This is identical to the return value of `which.packet`, which is a more accurate name.

`var.name`	vector of character strings or expressions as long as the number of conditioning variables. The contents are interpreted as names for the conditioning variables. Whether they are shown on the strip depends on the values of `strip.names` and `style` (see below). By default, the names are shown for shingles, but not for factors.
`factor.levels`	vector of character strings or expressions giving the levels of the conditioning variable currently being drawn. For more than one conditioning variable, this will vary with `which.given`. Whether these levels are shown on the strip depends on the values of `strip.levels` and `style` (see below). `factor.levels` may be specified for both factors and shingles (despite the name), but by default they are shown only for factors. If shown, the labels may optionally be abbreviated by specifying suitable components in `par.strip.text` (see `xyplot`)
`shingle.intervals`	if the current strip corresponds to a shingle, this should be a 2-column matrix giving the levels of the shingle. (of the form that would be produced by printing `levels(shingle)`). Otherwise, it should be `NULL`
`strip.names`	a logical vector of length 2, indicating whether or not the name of the conditioning variable that corresponds to the strip being drawn is to be written on the strip. The two components give the values for factors and shingles respectively. This argument is ignored for a factor when `style` is not one of 1 and 3.
`strip.levels`	a logical vector of length 2, indicating whether or not the level of the conditioning variable that corresponds to the strip being drawn is to be written on the strip. The two components give the values for factors and shingles respectively.
`sep`	character or expression, serving as a separator if the name and level are both to be shown.
`style`	integer, with values 1, 2, 3, 4 and 5 currently supported, controlling how the current level of a factor is encoded. Ignored for shingles (actually, when `shingle.intervals` is non-null. The best way to find out what effect the value of `style` has is to try them out. Here is a short description: for a style value of 1, the strip is colored in the background color with the strip text (as determined by other arguments) centered on it. A value of 3 is the same, except that a part of the strip is colored in the foreground color, indicating the current level of the factor. For styles 2 and 4, the part corresponding to the current level remains colored in the foreground color, however, for style = 2, the remaining part is not colored at all, whereas for 4, it is colored with the background color. For both these, the names of all the levels of the factor are placed on the strip from left to right. Styles 5 and 6 produce the same effect (they are subtly different in S, this implementation corresponds to 5), they are similar to style 1, except that the strip text is not centered, it is instead positioned according to the current level. Note that unlike S-PLUS, the default value of `style` is 1. `strip.names` and `strip.levels` have no effect if `style` is not 1 or 3.
`horizontal`	logical, specifying whether the labels etc should be horizontal. `horizontal=FALSE` is useful for strips on the left of panels using `strip.left=TRUE`
`par.strip.text`	list with parameters controlling the text on each strip, with components `col`, `cex`, `font`, etc.
`bg`	strip background color.
`fg`	strip foreground color.
`...`	arguments to be passed on to `strip.default`, overriding whatever value it would have normally assumed

Details

default strip function for trellis functions. Useful mostly because of the style argument — non-default styles are often more informative, especially when the names of the levels of the factor x are small. Traditional use is as strip = function(...) strip.default(style=2,...), though this can be simplified by the use of strip.custom.

Value

strip.default is called for its side-effect, which is to draw a strip appropriate for multi-panel Trellis conditioning plots. strip.custom returns a function that is similar to strip.default, but with different defaults for the arguments specified in the call.

Author(s)

Deepayan Sarkar Deepayan.Sarkar@R-project.org

Examples

## Traditional use
xyplot(Petal.Length ~ Petal.Width | Species, iris,
       strip = function(..., style) strip.default(..., style = 4))

## equivalent call using strip.custom
xyplot(Petal.Length ~ Petal.Width | Species, iris,
       strip = strip.custom(style = 4))

xyplot(Petal.Length ~ Petal.Width | Species, iris,
       strip = FALSE,
       strip.left = strip.custom(style = 4, horizontal = FALSE))

[Package lattice version 0.20-38 Index]