arow: Row Numbers Within the Augmented Row-Column Grid for a Table

Description

Return the row numbers associated with a specified table part or element, or with a set of row header values, within the augmented row-column grid of a table.

Usage

arow(x, id=NULL, hpath=NULL)

Value

A numeric vector containing row numbers within the table's augmented row-column grid. The row numbers are those partially or completely occupied by the cells associated with id or hpath. They will be increasing but not necessarily consecutive.

The returned vector may have length 0 if id refers to a table part or block that spans no rows, or if hpath matches no set of row header values.

Note that for a horizontal hvrule (hrule), the "row number" is actually a half-integer, bracketed by the table row numbers between which the hrule runs. For example, if the hrule runs between table rows 3 and 4, the returned value will be c(3.5).

Arguments

x: A textTable or a plotted table (pltdTable) object.
id: Character scalar containing the ID of a single table part, block, entry, or hvrule. (If x is a textTable, only the ID of a table part is allowed.)
hpath: Character vector with length between 0 and the number of layers in the row header. The i-th element should be one of the values in the i-th row header column, or NA. See DETAILS. Only one of id and hpath should be specified.

Details

See the documentation for adim for more information about the augmented row-column grid of a table.

Only one of arguments id and hpath should be specified. id is searched for first among table parts (the only thing available for a textTable), then blocks, entries, and hvrules, in that order. The search stops at the first match. It is an error if id is not found in any of these.

hpath is short for "header path". It is used to obtain row numbers associated with specified combinations of values of the row header variables. Suppose there are L layers of row headers. If the length of hpath is less than L, NA values are added at the end to reach that length. The function returns the intersection of the row numbers for which the i-th outermost of the header layers equals the i-th element of hpath. An NA in hpath is taken to match all values in the correponding layer of row headers. Thus, if L == 4 and hpath=c("a", NA, "c"), the function will return the row numbers for which the outermost row header has a value of "a" _and_ the third outermost has a value of "c". If no row has the combination of values specified by hpath then the returned vector will have length 0.

Since hpath refers to values of row header variables, it cannot be used to get row numbers associated with table annotation, or with the column header. (Use id instead.)

Examples

Run this code

ttbl <- textTable(iris2_tab, title=c("Title 1", "2nd title"), foot="Foot")
plt <- plot(ttbl)

arow(plt, id="title")  # block "title" spans first two rows
arow(plt, id="body,4,2")  # single entry
# Remove the first line of the column header:
plot(ttbl[-arow(ttbl, id="colhead")[1], ])
# Remove the "versicolor" species (a value in row header layer 1):
plot(ttbl[-arow(ttbl, hpath=c("versicolor")), ])
# Remove the means for all species (a value in row header layer 2):
plot(ttbl[-arow(ttbl, hpath=c(NA, "mean")), ])
# Remove the mean just for the versicolor species:
plot(ttbl[-arow(ttbl, hpath=c("versicolor", "mean")), ])