rhandsontable-table

name	rhandsontable-table
description	Auto-invoked when creating or modifying rhandsontable tables in Shiny. Governs rHandsontableOutput/renderRHandsontable patterns, reading edits back to R, column types and validation, read-only controls, and hot_to_r() usage.

rhandsontable Table Skill

Canonical Shiny Pattern

library(rhandsontable)

# --- UI ----------------------------------------------------------------------
rHandsontableOutput("my_table")

# --- Server ------------------------------------------------------------------
output$my_table <- renderRHandsontable({
  req(data())
  rhandsontable(data(), rowHeaders = NULL) %>%
    hot_table(highlightCol = TRUE, highlightRow = TRUE)
})

# Read edits back into R
edited_data <- reactive({
  req(input$my_table)
  hot_to_r(input$my_table)
})

hot_to_r() converts the Handsontable JSON input back to an R data frame. Always use hot_to_r() — do not parse input$tableId manually.

---

Column Configuration with hot_col()

rhandsontable(data()) %>%
  hot_col("status",    type = "dropdown", source = c("Active", "Completed", "Withdrawn")) %>%
  hot_col("confirmed", type = "checkbox") %>%
  hot_col("visit_dt",  type = "date",     dateFormat = "YYYY-MM-DD") %>%
  hot_col("aval",      type = "numeric",  format = "0.00") %>%
  hot_col("notes",     type = "text")

Column types: "text", "numeric", "date", "checkbox", "dropdown", "autocomplete", "password", "time", "handsontable".

---

Read-Only Columns

# Make specific columns read-only
rhandsontable(data()) %>%
  hot_col("usubjid", readOnly = TRUE) %>%
  hot_col("studyid", readOnly = TRUE)

# Make the entire table read-only
rhandsontable(data(), readOnly = TRUE)

---

Column Widths and Visibility

rhandsontable(data()) %>%
  hot_col("usubjid", width = 120) %>%
  hot_col("notes",   width = 250) %>%
  hot_cols(columnSorting = TRUE)    # enable sorting on all columns

# Hide a column (keep data, hide visually)
rhandsontable(data()) %>%
  hot_col("internal_id", width = 0.1)   # near-zero width to hide

---

Table-Level Options

rhandsontable(
  data(),
  rowHeaders   = NULL,       # hide row numbers
  stretchH     = "all",      # stretch columns to fill width
  contextMenu  = TRUE,       # right-click menu
  overflow     = "hidden",   # clip overflow
  height       = 400,        # fixed height with scroll
  width        = "100%"
) %>%
  hot_table(
    highlightCol  = TRUE,
    highlightRow  = TRUE,
    enableComments = FALSE
  )

---

Conditional Cell Formatting

# Use hot_cols() with renderer for custom cell styling
rhandsontable(data()) %>%
  hot_col(
    "aval",
    renderer = "
      function(instance, td, row, col, prop, value, cellProperties) {
        Handsontable.renderers.NumericRenderer.apply(this, arguments);
        if (value > 100) {
          td.style.background = '#f8d7da';
          td.style.color = '#721c24';
        }
      }
    "
  )

Renderer strings are JavaScript. Keep them simple — complex logic belongs in R before the table renders.

---

Preserving Edits Across Re-Renders

The table re-renders whenever renderRHandsontable re-executes, losing unsaved edits. Preserve edits by seeding the render from the input:

# Seed table from edited state if available, else from source data
output$my_table <- renderRHandsontable({
  current <- if (!is.null(input$my_table)) hot_to_r(input$my_table) else data()
  rhandsontable(current, rowHeaders = NULL)
})

Use this pattern when the table data can also change from outside (e.g., a reset button). Add an ignoreInit = TRUE guard if needed to avoid circular reactivity.

---

Validation

rhandsontable(data()) %>%
  hot_col(
    "age",
    type   = "numeric",
    validator = "function(value, callback) { callback(value >= 0 && value <= 120); }",
    allowInvalid = FALSE   # blocks invalid input (shows red border)
  )

allowInvalid = TRUE (default) highlights invalid cells but still accepts them. Set FALSE to block the entry.

---

Save / Reset Pattern in Shiny

# UI
rHandsontableOutput("tbl")
actionButton("save",  "Save Changes")
actionButton("reset", "Reset")

# Server
rv <- reactiveValues(data = initial_data)

output$tbl <- renderRHandsontable({
  rhandsontable(rv$data, rowHeaders = NULL)
})

observeEvent(input$save, {
  rv$data <- hot_to_r(input$tbl)
  # write rv$data to DB or file here
})

observeEvent(input$reset, {
  rv$data <- initial_data
})

---

Common Pitfalls

• Always use hot_to_r() — input$tableId is a JSON list, not a data frame; hot_to_r() converts it correctly
• Re-render wipes edits — any reactive dependency in renderRHandsontable that changes will discard unsaved edits; seed from input$tableId or use reactiveValues to persist
• Column order must match data frame — hot_col() refers to columns by name; if you reorder columns in the data frame, verify hot_col() calls still target the right columns
• Date columns — rhandsontable returns dates as strings; coerce with as.Date() after hot_to_r()
• Large data — Handsontable renders all rows in the DOM; avoid > 500 rows; use pagination or aggregation upstream
• Circular reactivity — if renderRHandsontable reads input$my_table to seed itself, wrap with isolate() or use reactiveValues to break the cycle