Choices for a categorical colour, shape, group or fill selector

The columns facet_check() accepts plus the logical columns, in data order, for any selector that maps one colour, shape or group to each level of a column. A column of IDs or gene names would ask for one level per row, so categorical columns with max.levels or more distinct values are left out, while a TRUE/FALSE column, which facet_check() leaves out of the split choices, is a sensible shape or group. Every built-in module builds its categorical Color By, Shape By, Group By and Fill By choices this way.

Usage

discrete_choices(data, max.levels = 50, keep = NULL)

Arguments

data

A data frame.

max.levels

A single positive number, passed to facet_check(). Categorical columns with this many or more distinct values are excluded.

keep

Column names to keep whatever their size, such as the column a caller named in defaults, so an explicit choice is always on offer. Names not in data are ignored.

Value

A character vector of column names, in the order they appear in data. character(0) when data is NULL or has no columns.

Details

A selector that can also draw a numeric column as a gradient (the scatter module's Color By, say) should add its numeric columns to these itself.

See also

facet_check() for facet and split selectors.

Author

Jared Andrews

Examples

df <- data.frame(
    sample = paste0("S", 1:60),
    group = rep(c("A", "B", "C"), 20),
    treated = rep(c(TRUE, FALSE), 30),
    score = rnorm(60)
)
# `sample` has 60 levels, so only `group` and `treated` are offered.
discrete_choices(df)
#> [1] "group"   "treated"

# A column named in `defaults` is offered regardless.
defaults <- list(group.by = "sample")
discrete_choices(df, keep = get_default(defaults, "group.by", NULL))
#> [1] "sample"  "group"   "treated"