Skip to contents

Modify column merging

Source: R/modify_column.R
modify_column_merge.Rd

Merge two or more columns of a table into one. Run show_header_names() to see the column names.

Usage

modify_column_merge(x, pattern, rows = NULL)

remove_column_merge(x, columns = everything())

Arguments

x

(ltsummary)
An ltsummary object.

pattern

(string)
A pattern with column names of x$table_body in curly braces, e.g. "{conf.low}, {conf.high}" to build a confidence interval. The merged text is shown in the first column named in the pattern; the other columns are hidden.

rows

(predicate expression)
Expression selecting rows of x$table_body, e.g. row_type == "label". Used for footnotes, formatting functions, missing symbols, indentation and text formatting. Default is NULL, which selects every row (or, for a footnote, places it in the column header). See the section below.

columns

(selector)
Columns of x$table_body.

Value

An ltsummary object.

Details

  1. Calling this function only records the instruction to merge. The merging happens when the table is printed or converted with as_lt() or as.data.frame.ltsummary(), after the formatting functions have been applied to each column, so modify_fmt_fun() still works on the individual columns.

  2. remove_column_merge() undoes merges recorded with modify_column_merge(). It also undoes merges made by the package, for example the "{label}, {stat_label}" merge of add_stat_label().

rows argument

The rows argument is a predicate expression that selects the rows the styling applies to. It is evaluated in x$table_body, so its column names are available as variables, and it must give a logical vector. To style the rows of the age variable pass rows = variable == "age"; to style the variable label rows pass rows = row_type == "label". A logical vector, a vector of row numbers or a function of table_body are also accepted.

Objects from the calling environment can be used in the expression, e.g. rows = variable == vars[1], as long as they do not share a name with a column of x$table_body. The expression is stored, not its result, and resolved when the table is printed, so it survives later changes to the table such as remove_row_type().

See also

Other advanced modifiers: modify_bold_italic, modify_column_alignment(), modify_column_hide(), modify_fmt_fun(), modify_indent(), modify_missing_symbol(), modify_post_fmt_fun(), tbl_butcher()

Examples

# Example 1 ----------------------------------
# show the test statistic next to the p-value
trial |>
  tbl_summary(by = trt, missing = "no", include = c(age, marker)) |>
  add_p(all_continuous() ~ "t.test", pvalue_fun = label_style_pvalue(prepend_p = TRUE)) |>
  modify_fmt_fun(statistic ~ label_style_sigfig()) |>
  modify_column_merge(pattern = "t = {statistic}; {p.value}") |>
  modify_header(statistic = "**t-test**")


# Example 2 ----------------------------------
# merge only in some rows
trial |>
  tbl_summary(by = trt, include = c(age, grade), missing = "no") |>
  add_p() |>
  modify_column_merge(pattern = "{stat_1} / {stat_2}", rows = row_type == "level") |>
  modify_header(stat_1 = "**Drug A / Drug B**")