Line up the decimal points of a numeric column

Marks one or more body columns to be rendered as two adjacent RTF cells: the part before the first decimal separator (right-aligned) and the separator plus everything after it (left-aligned). The decimal points then line up exactly, whatever the font, instead of merely right-aligning the last character.

Usage

set_decimal_split(x, ...)

# S3 method for class 'rtftable'
set_decimal_split(
  x,
  cols = NULL,
  ratio = NULL,
  decimal_mark = ".",
  pad_chars = c(1, 1),
  min_chars = c(4, 6),
  max_chars = 10,
  include_compound = FALSE,
  ...
)

# S3 method for class 'list'
set_decimal_split(x, ...)

Arguments

x: An rtftable(), or a list of them (pages from as_rtftables()).
...: Unused.
cols: Columns to split: names or positions. NULL clears the setting.
ratio: NULL (the default) to size the two halves automatically from the widest left and right part actually present, or a number in (0, 1) giving the left half's share of the column width. Pass an explicit value to keep the split point identical across pages, whose data differ.
decimal_mark: The separator to split at. Default ".".
pad_chars: Breathing room added to each measured half before the ratio is taken, in character-width units, integer half first. Default c(1, 1).
min_chars: Floor for each half, in the same units, applied after pad_chars: the effective width of a half is max(measured + pad_chars, min_chars). Default c(4, 6) – a three-digit integer part and a five-character decimal part, each plus its padding character – so anything at or below that size holds a 4 : 6 baseline and anything larger scales from its own measurement. Without a floor a column of 0.0000 values measures 1 against 5 and leaves the integer half cramped. Pass pad_chars = c(0, 0), min_chars = c(0, 0) for the raw measured ratio.
max_chars: Total measured width, in the same units, above which the allowance is dropped and the raw measured proportions are used however lopsided. Default 10: a column that long needs every twip it has, and reserving room it does not use would squeeze the digits it does. Inf never drops the allowance.
include_compound: Split values carrying a whitespace- or parenthesis-separated companion ("12.3 (4.56)") too. Default FALSE.

Value

An object of the same shape as x.

Details

This is an RTF-output option, not a data transformation. The table's data frame is never rewritten, the total table width is unchanged (only the column's own width is divided), and the column-header block keeps rendering over the original single column. The console preview (print() / format()) therefore does not show the split.

Which cells are split

A cell is split when it starts with an optional relational or sign prefix (<, >, >=, <=, ~, +, -) followed by a digit or the separator, and contains at least one digit. The prefix travels with the left half and any suffix (%, a ^{a} footnote marker, ...) with the right half, so both hang outside the aligned point:

cell	left	right
`3.45`	`3`	`.45`
`-0.7`	`-0`	`.7`
`<0.001`	`<0`	`.001`
`45.6%`	`45`	`.6%`
`100`	`100`	(empty)

Everything else – free text such as "n (%)" or a group label – is rendered as one cell across the pair, exactly as it looks without the option. Leading and trailing spaces (non-breaking ones included, as left by fmt_right_align()) are dropped from split cells, since the split supplies the alignment they were emulating.

Compound values such as "12.3 (4.56)" (mean (SD)) are excluded by default: the trailing group would dominate the right half and drag the split point far to the left. Set include_compound = TRUE to split them anyway.

A selected column in which no cell carries a separator is left untouched. That decision is made per table, so across paginated pages a column may be split on one page and not on another – a PK visit column that is all BLQ in one time band, for instance. Each page stays internally consistent, which is what the geometry needs.

Examples

df  <- data.frame(
  Statistic = c("Mean", "SD", "p-value"),
  Value     = c("12.3", "1.05", "<0.001"),
  stringsAsFactors = FALSE
)
tbl <- rtftable(df) |> set_decimal_split(cols = "Value")