Skip to contents

Bind and format domain-specific tables or a single table

Source: R/margot_bind_tables.R
margot_bind_tables.Rd

This function binds a named list of domain-specific tables, or handles a single data frame. It optionally sorts by the E-value bound column, renames columns, highlights rows above a threshold, and groups rows by domain when provided. Works for latex, html, or markdown output. It accepts both `transformed_table` layouts of [margot_plot()]: the default reporting layout, whose bound column is "E-value bound" and whose outcomes are in an "Outcome" column, and the legacy layout, whose bound column is "E_Val_bound" and whose outcomes are row names. A table that already has an "Outcome" column keeps it, and row names are not printed.

Usage

margot_bind_tables(
  tables_list,
  sort_E_val_bound = c("none", "asc", "desc"),
  e_val_bound_threshold = 1.1,
  highlight_color = "yellow",
  bold = TRUE,
  output_format = "markdown",
  rename_cols = TRUE,
  col_renames = list(`E-Value` = "E_Value", `E-Value bound` = "E_Val_bound"),
  rename_ate = FALSE,
  threshold_col = "E_Val_bound",
  kbl_args = list(booktabs = TRUE, caption = "Combined Results", align = NULL)
)

Arguments

tables_list

a named list of data frames, each a domain table, or a single data frame. For lists, row names are captured as an "Outcome" column; for data frames, same applies.

sort_E_val_bound

character: one of "none", "asc", or "desc". default "none". For a list of domain tables, rows are sorted within each domain and the domains keep their order. Every table in a list must have the same columns once its outcome and E-value bound columns take the first table's spelling and any difference in interval level is resolved; otherwise binding would split values across columns, and the function stops with an error naming the differing columns.

Tables that differ only in their interval level, for example domains each Bonferroni-corrected over their own outcomes ("99.58% CI lower" in one, "99.17% CI lower" in another), bind with level-free headers ("CI lower", "CI upper", "CI relative to 0" and the original-unit equivalents) and a per-row "CI level" column (for example "99.58%") after the last interval limit. Legacy tables with different quantile headers (for example "0.21 %" and "0.42 %") take "CI lower" and "CI upper", and their `confidence_level` column keeps each row's level. Tables that share one level keep their headers unchanged.

Tables from the report and legacy layouts of margot_plot(), or with different estimate columns, renames or E-value thresholds, cannot be bound together.

e_val_bound_threshold

numeric; threshold for highlighting rows. default 1.1.

highlight_color

character or NULL; background colour for highlighted rows. default "yellow".

bold

logical; whether to bold highlighted rows. default TRUE.

output_format

character; one of "latex", "html", or "markdown". default "markdown".

rename_cols

logical; whether to apply column renaming. default TRUE.

col_renames

named list; mapping new_name = old_name for renaming. default list("E-Value" = "E_Value", "E-Value bound" = "E_Val_bound").

rename_ate

logical; if TRUE, renames the effect column to "ATE". default FALSE.

threshold_col

character; name of the (pre-rename) column to test and sort by. default "E_Val_bound". When that column is absent, the first of "E_Val_bound", "E-value bound", "E-Value bound" and "E-Value Bound" that is present is used.

kbl_args

list; extra arguments for kableExtra::kbl(). default list(booktabs=TRUE, caption="Combined Results", align=NULL).