safePivot

safePivot is an R htmlwidget for safer interactive drag-and-drop pivot tables. It wraps PivotTable.js and adds R-focused safeguards for exploratory data analysis, Shiny export, and reproducible reporting workflows.

The main idea is simple:

1. use the browser for interactive pivot-table exploration;
2. capture the current pivot configuration in Shiny;
3. recompute the same pivot result in R;
4. export the R-computed result to CSV, Excel, RDS, or RData.

safePivot is especially useful when empty cells, observed missing values, and observed zero values must stay clearly separated.

Installation

Install the development version from GitHub:

# install.packages("remotes")
remotes::install_github("dongwenluo/safePivot")

Load the package:

library(safePivot)

Key features

• Interactive PivotTable.js drag-and-drop UI.
• Safe R-side pivot computation with safe_pivot_compute().
• Shiny config capture with input$<outputId>_config.
• Config-based export with safe_pivot_compute_from_config().
• CSV, Excel, RDS, and RData export helpers.
• Explicit missing, non-missing, zero, and non-zero aggregators.
• Within-cell, row, column, and total percentage aggregators.
• Sum/count fraction aggregators with explicit denominators.
• R factor-level ordering.
• Data-type badges for draggable variables.
• Browser-side conditional formatting for table output.
• Starting filters (inclusions / exclusions) that are also applied on export.
• Unused variables listed on the left for wide data (unused_layout).
• Heatmap renderers with selectable palettes.
• Lightweight Plotly chart renderers.

Basic use

safePivot(
  iris,
  rows = "Species",
  vals = "Sepal.Length",
  aggregator = "Median",
  renderer = "Table"
)

Drag variables between the unused list, the rows area (left) and the columns area (top); pick the aggregator and value variable at top left.

Heatmap renderer

safePivot(
  mtcars,
  rows = "cyl",
  cols = "gear",
  vals = "mpg",
  aggregator = "Mean",
  renderer = "Heatmap",
  heatmap_palette = "green"
)

Available heatmap palettes are:

"blue"
"yellow_orange"
"green"
"blue_white_red"
"green_red"

Available heatmap renderers are:

"Heatmap"
"Row Heatmap"
"Col Heatmap"

Browser-side conditional formatting is intentionally disabled for heatmap renderers so that heatmap palettes are not overwritten by SafePivot table-cell formatting classes.

Lightweight Plotly charts

safePivot includes lightweight Plotly renderers for quick visual summaries:

safePivot(
  mtcars,
  rows = "cyl",
  cols = "gear",
  vals = "mpg",
  aggregator = "Mean",
  renderer = "Bar Chart",
  plot_default_height = 520
)

Supported chart renderers include:

"Horizontal Bar Chart"
"Horizontal Stacked Bar Chart"
"Bar Chart"
"Stacked Bar Chart"
"Line Chart"
"Area Chart"
"Scatter Chart"
"Multiple Pie Chart"

Treemap is intentionally not included in the core package because it requires a larger Plotly bundle and extra hierarchy-specific renderer logic.

Data-type badges

safePivot can show compact badges beside draggable variables. These badges help users decide which variables are suitable for rows, columns, or values before they start dragging.

Badge	Meaning
num	numeric / double
int	integer
chr	character
fct	factor
ord	ordered factor
date	Date
time	POSIX date-time
lgl	logical
list	list column
obj	other object type

badge_dat <- data.frame(
  id = 1:3,
  score = c(1.2, 3.4, 5.6),
  group = factor(c("A", "B", "A")),
  ordered_group = factor(
    c("low", "medium", "high"),
    levels = c("low", "medium", "high"),
    ordered = TRUE
  ),
  name = c("x", "y", "z"),
  flag = c(TRUE, FALSE, TRUE),
  harvest_date = as.Date(c("2026-01-01", "2026-01-02", "2026-01-03")),
  timestamp = as.POSIXct(c(
    "2026-01-01 08:00:00",
    "2026-01-02 08:00:00",
    "2026-01-03 08:00:00"
  ))
)

safePivot(
  badge_dat,
  vals = "score",
  aggregator = "Mean",
  show_type_badges = TRUE,
  unused_layout = "top"
)

Factor order

If a variable is an R factor, safePivot can respect its factor-level order.

iris2 <- iris
iris2$Species <- factor(
  iris2$Species,
  levels = c("virginica", "versicolor", "setosa")
)

safePivot(
  iris2,
  rows = "Species",
  vals = "Sepal.Length",
  aggregator = "Median",
  renderer = "Table",
  respect_factor_order = TRUE
)

For ordered factors, define levels explicitly instead of relying on alphabetical order:

ordered_dat <- data.frame(
  stage = factor(
    c("low", "medium", "high"),
    levels = c("low", "medium", "high"),
    ordered = TRUE
  ),
  value = c(1, 2, 3)
)

Aggregation rules

safePivot uses explicit aggregation rules so that empty pivot cells, observed missing values, and observed zero values are not confused.

This is the beginning of the README. Read the rest at the original. Read the full README ↗