datasetviewer

An interactive, SAS Studio-style dataset viewer for R. datasetviewer renders a dataset in a fast, scrollable grid modelled on the SAS Studio table viewer, from one htmlwidget that runs in interactive Shiny apps and in static HTML documents.

Fast on large data, no sampling. The data is sent to the browser as Parquet and queried in place with DuckDB-WASM, so filtering, sorting, and scrolling stay responsive on datasets that overwhelm DOM-bound grids or server-paged tables.
SAS Studio layout. A column-selection panel with char/num type icons and a collapse toggle, a property pane (Label, Name, Length, Type, Format), a names-versus-labels header toggle, click-a-header to sort (Shift-click for multi-column sort), and a free-text “Filter Table Rows” expression.
Clinical metadata. With the sibling artoo package installed, column labels, formats, and storage lengths are read from a labelled or CDISC-conformed frame. Plain data frames work with zero artoo dependency.
Runs offline / behind a firewall. The grid, filters, code view, and CSV export are bundled in the package. The DuckDB-WASM engine loads from a CDN by default but is fetched into the package at install time, so a Shiny app can serve it to browsers with no internet at runtime.

Installation

Install the released version from CRAN:

install.packages("datasetviewer")

Or the development version from GitHub:

# install.packages("pak")
pak::pak("vthanik/datasetviewer")
# or
remotes::install_github("vthanik/datasetviewer")

Usage

library(datasetviewer)

# a plain data frame
dataset_viewer(mtcars)

# an artoo-conformed frame, headers shown as labels
adsl <- artoo::read_dataset("adsl.parquet")
dataset_viewer(adsl, view = "labels")

Click a column header to select it, then click again to cycle its sort — ascending, descending, then back to unsorted. Shift-click additional headers to build a multi-column sort, each shown with its direction and priority (AGE ↑1, SEX ↓2). In the right-click menu, Sort Ascending/Descending add that column to the sort and Clear Sorting removes just that column; you can also add a filter or copy the column or its header. Click the funnel in the toolbar to edit the filter expression directly (for example AGE > 50 and SEX = "M").

In Shiny

library(shiny)
ui <- fluidPage(datasetviewerOutput("viewer", height = "500px"))
server <- function(input, output, session) {
  output$viewer <- renderDatasetViewer(dataset_viewer(adsl))
  # input$viewer_filter, input$viewer_sort, input$viewer_columns,
  # input$viewer_view track the user's current view
}
shinyApp(ui, server)

Offline and corporate use

Everything except the DuckDB-WASM query engine (~35 MB, too large to ship in a package) is bundled and works offline. The engine loads from a CDN by default, but — like the arrow package’s C++ library — it is fetched into the package at install time, so a Shiny app can serve it to browsers with no internet at runtime. Steer the install-time fetch for air-gapped or mirrored environments:

Variable	Effect
`DATASETVIEWER_DUCKDB_DIR`	Copy the engine from a pre-staged directory (fully air-gapped install).
`DATASETVIEWER_DUCKDB_URL` / `DATASETVIEWER_DUCKDB_EXT_URL`	Internal mirror base URLs for the engine and the DuckDB extension repository.
`DATASETVIEWER_DUCKDB_OFFLINE`	Set to `true` to skip the fetch and always use the CDN.

The install never fails if the engine cannot be fetched; the widget falls back to the CDN at runtime. For self-contained HTML documents, set options(datasetviewer.use_local_engine = FALSE) to load the engine from the CDN instead of embedding it. See vignette("datasetviewer") for details.