---
title: "Referential Footnotes"
author: "Emily de la Rua"
date: "`r Sys.Date()`"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Referential Footnotes}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
suggested_dependent_pkgs <- c("dplyr")
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  eval = all(vapply(
    suggested_dependent_pkgs,
    requireNamespace,
    logical(1),
    quietly = TRUE
  ))
)
```

---------

## Introduction

There is currently no formal method for adding referential footnotes to a listing object. This vignette demonstrates how referential footnotes can be added to a `listing_df` object via a workaround applied during pre-processing.

To learn more about how listings are constructed using the `rlistings` package, see the [Getting Started vignette](rlistings.html).

---------

## Referential Footnotes Workaround

When creating a listing with the `rlistings` package, you may want to add referential footnotes, similar to how referential footnotes can be added to `rtable` objects. Since there is no formal method in `rlistings` for applying referential footnotes to a `listing_df` object, we will demonstrate how a workaround can be applied to add a set of pseudo-referential footnotes to your listing.

To demonstrate, we will create a basic listing below.

We begin by loading in the `rlistings` package.

```{r, message=FALSE}
library(rlistings)
library(dplyr)
```

For this example, we will use the dummy ADAE dataset provided within the `formatters` package as our data frame, which consists of 48 columns of adverse event patient data, and one or more rows per patient. For the purpose of this example, we will subset the data and only use the first 30 records of the dataset.

```{r}
adae <- ex_adae[1:30, ]
```

Now we will create a basic listing.

```{r}
lsting <- as_listing(
  df = adae,
  key_cols = c("ARM", "USUBJID", "ASEQ", "ASTDY"),
  disp_cols = c("BMRKR1", "AESEV"),
)

lsting
```

For this example, we will add 4 referential footnotes.

1. In the `ARM` column, for all records with `ARM = "A: Drug X"` and `ASEQ` equal to 1 or 2.
2. In the `ASTDY` column, for imputed dates, where imputed dates are indicated by the `ASTDTF` variable.
3. In the `AESEV` column, for all records with `AETOXGR` equal to 5.
3. In the `USUBJID` column header.

Footnote text can be supplied either as a vector, where each element is a new footnote, or as a single string with footnotes separated by the new line (`\n`) character. For example, see the following list of referential footnotes:

```{r}
ref_fns <- "*   ASEQ 1 or 2\n**  Analysis start date is imputed\n***  Records with ATOXGR = 5\n**** ID column"
```

We start with our first footnote in the `ARM` column for records with arm "A: Drug X" which also have analysis sequence number 1 or 2. Referential footnotes can be added to any variable by converting the variable to a factor, editing the variable values, and adding a corresponding footnote below the listing. We will add our first referential footnote in the `ARM` column. To ensure that levels are correctly ordered, be sure to specify the new level order when mutating your variable.

```{r}
# Save variable labels for your data to add back in after mutating dataset
df_lbls <- var_labels(adae)

# Mutate variable where referential footnotes are to be added according to your conditions
# Specify order of levels with new referential footnotes added
adae <- adae %>% dplyr::mutate(
  ARM = factor(
    ifelse(ARM == "A: Drug X" & ASEQ %in% 1:2, paste0(ARM, "*"), as.character(ARM)),
    levels = c(sapply(levels(adae$ARM), paste0, c("", "*")))
  )
)

# Add data variable labels back in
var_labels(adae) <- df_lbls

# Generate listing
lsting <- as_listing(
  df = adae,
  key_cols = c("ARM", "USUBJID", "ASEQ", "ASTDY"),
  disp_cols = c("BMRKR1", "AESEV"),
)

lsting
```

Additional referential footnotes can be added to the data be repeating the above steps.

For example, we can add the second referential footnote to the `ASTDY` column for imputed analysis start days. We use dummy variable `ASTDTF` to indicate imputed analysis start dates. When adding referential footnotes to numeric variables, the variables must be converted to factors.

```{r}
set.seed(1)

# Save variable labels for your data to add back in after mutating dataset
df_lbls <- var_labels(adae)

# Mutate variable where referential footnotes are to be added according to your conditions
# Specify order of levels with new referential footnotes added
adae <- adae %>%
  dplyr::mutate(ASTDTF = sample(c("Y", NA), nrow(.), replace = TRUE, prob = c(0.25, 0.75))) %>%
  dplyr::mutate(ASTDY = factor(
    ifelse(!is.na(ASTDTF), paste0(as.character(ASTDY), "**"), as.character(ASTDY)),
    levels = c(sapply(sort(unique(adae$ASTDY)), paste0, c("", "**")))
  )) %>%
  dplyr::select(-ASTDTF)

# Add data variable labels back in
var_labels(adae) <- df_lbls

# Generate listing
lsting <- as_listing(
  df = adae,
  key_cols = c("ARM", "USUBJID", "ASEQ", "ASTDY"),
  disp_cols = c("BMRKR1", "AESEV"),
)

lsting
```

Next we can add our third referential footnote to the `AESEV` column for records with analysis toxicity grade 5.

```{r}
# Save variable labels for your data to add back in after mutating dataset
df_lbls <- var_labels(adae)

# Mutate variable where referential footnotes are to be added according to your conditions
# Specify order of levels with new referential footnotes added
adae <- adae %>% dplyr::mutate(
  AESEV = factor(
    ifelse(AETOXGR == 5, paste0(AESEV, "***"), as.character(AESEV)),
    levels = c(sapply(levels(adae$AESEV), paste0, c("", "***")))
  )
)

# Add data variable labels back in
var_labels(adae) <- df_lbls

# Generate listing
lsting <- as_listing(
  df = adae,
  key_cols = c("ARM", "USUBJID", "ASEQ", "ASTDY"),
  disp_cols = c("BMRKR1", "AESEV"),
)

lsting
```

Referential footnotes can also be added to column header labels. Suppose we want to add our fourth referential footnote to the `USUBJID` column label. We can do so by editing the `USUBJID` variable label and adding the footnote text as follows:

```{r}
# Modify data variable label
adae <- adae %>% var_relabel(
  USUBJID = paste0(var_labels(adae)[["USUBJID"]], "****")
)

# Generate listing
lsting <- as_listing(
  df = adae,
  key_cols = c("ARM", "USUBJID", "ASTDY", "ASEQ"),
  disp_cols = c("BMRKR1", "AESEV")
)

lsting
```

Finally, we add in the referential footnote text below the listing as follows.

```{r}
# Generate listing
lsting <- as_listing(
  df = adae,
  key_cols = c("ARM", "USUBJID", "ASTDY", "ASEQ"),
  disp_cols = c("BMRKR1", "AESEV")
)

main_footer(lsting) <- c(main_footer(lsting), ref_fns)

lsting
```

Now our listing is complete, with all four referential footnotes denoted within the listing and described in the footnotes section below.

---------

## Summary

In this vignette, you have learned how a pre-processing workaround can be applied to add referential footnotes to listings.

**For more information please explore the [rlistings website](https://insightsengineering.github.io/rlistings/main/).**