jupyterlab-tabular-data-viewer-extension 1.6.12

Jupyterlab extension to browse tabular data files (Parquet, Excel, CSV, TSV) with filtering and sorting capabilities

jupyterlab_tabular_data_viewer_extension

[!TIP] This extension is part of the stellars_jupyterlab_extensions metapackage. Install all Stellars extensions at once: pip install stellars_jupyterlab_extensions

View and browse Parquet, Excel, CSV, and TSV files directly in JupyterLab. Double-click any .parquet, .xlsx, .csv, or .tsv file to open it in a simple, spreadsheet-like table view - no code required (yes, really). Navigate through your data, inspect values, and explore the structure of your tabular data files with interactive column resizing and advanced filtering capabilities.

Full disclosure: This is a shameless ripoff of your typical tabular data browsing tools. Zero ingenuity, zero creativity - just unabashed borrowing of ideas that worked elsewhere. If it looks familiar, that's the point.

Opening files: Right-click any supported file and select "Tabular Data Viewer" from the "Open With" menu, or simply double-click to open with the default viewer.

Column statistics: Hover over any column header to reveal an info icon, click it to view comprehensive statistics.

Context menu: Right-click any row to copy data as JSON.

Export: Click the Export link in the status bar (or right-click on the viewer) to export the current view in your choice of format - original, Excel (.xlsx), CSV, Parquet (.parquet), or JSONL (.jsonl). When filters are active, the export popup notes that only filtered rows will be exported.

Features

Supported File Formats:

• Parquet files (.parquet) - Full support with efficient columnar data reading
• Excel files (.xlsx) - Multi-sheet support: a sheet bar appears at the bottom for workbooks with more than one sheet, and switching sheets resets all filters/sort/selection (each sheet behaves like a separate file). Mixed-type columns (e.g. integers and strings in the same column) are handled via per-column cascading type inference rather than failing to open. Excel files must still be simple tabular data without merged cells, complex formulas, or advanced formatting
• CSV files (.csv) - Comma-separated values with UTF-8 encoding (fallback to latin1)
• TSV files (.tsv) - Tab-separated values with UTF-8 encoding (fallback to latin1)

Core viewing and navigation:

• Simple table display showing your data in familiar spreadsheet format
• Column headers with field names and simplified datatype indicators
• Interactive column resizing - drag column borders to adjust width independently
• Frozen index column - row numbers stay fixed when scrolling horizontally through wide datasets
• Row selection - click anywhere on a row to highlight it with subtle color shading. Click again to deselect, or click another row to switch selection
• Progressive loading - starts with 500 rows, automatically loads more as you scroll (your patience rewarded)
• File statistics (column count, row count, file size) at a glance
• Fixed status bar remains visible during horizontal scrolling (because it got tired of moving)
• Handles large files efficiently with server-side processing

Advanced filtering and sorting:

• Column sorting with three-state toggle (ascending, descending, off)
• Per-column filtering with substring or regex pattern matching
• Multi-select value filter - Click filter button next to any column to select from unique values with counts. Supports filtering on empty strings and null values
• Case-insensitive search option
• Numerical filters supporting comparison operators (>, <, >=, <=, =)
• Clear filters functionality to reset all active filters
• Multiple filters work together to narrow down results

Additional features:

• Column statistics modal - View comprehensive statistics including data type, row counts, null values, unique counts, and type-specific metrics (numeric: min/max/mean/median/std dev/outliers; string: most common value/length stats; date: earliest/latest dates). Includes scrollable list of unique values sorted by frequency with counts and percentages. Copy statistics as JSON with one click
• Export - Export link in the status bar (or right-click on the viewer) opens a format picker: original, Excel (.xlsx), CSV, Parquet (.parquet), or JSONL (.jsonl). Exports preserve active filters and sort order. Filename includes the slugified sheet name for multi-sheet Excel and a _filtered suffix when filters are applied
• Right-click context menu on rows to copy data as JSON
• Refresh view - Right-click on viewer and select "Refresh View" to reload data from file while preserving scroll position, filters, and sorting
• Cell text truncation - Configurable maximum character limit for cell display (default: 100 characters). Text longer than limit shows "..." ellipsis. Set to 0 for unlimited display
• Complex data types display - List/tuple and dict values display as JSON strings for easy inspection of nested/structured data
• Absolute row indices - Row numbers always show original file position, even with active filters or sorting
• Configurable file type support via Settings - Enable/disable Parquet, Excel, or CSV/TSV handling
• All features work seamlessly across all supported file formats

Installation

Requires JupyterLab 4.0.0 or higher.

pip install jupyterlab_tabular_data_viewer_extension

Uninstall:

pip uninstall jupyterlab_tabular_data_viewer_extension

Configuration

Configure extension behavior through JupyterLab Settings:

1. Open Settings → Settings Editor
2. Search for "Tabular Data Viewer Extension"
3. Configure options:
   • Enable Parquet files - Default: enabled
   • Enable Excel files - Default: enabled
   • Enable CSV files - Default: enabled
   • Enable TSV files - Default: enabled
   • Maximum Cell Characters - Default: 100. Maximum characters to display in a cell before truncating with "...". Set to 0 for unlimited display
   • Maximum Unique Values - Default: 100. Maximum number of unique values to display in filter dialog and column statistics. Set to 0 for no limit

When a file type is disabled, files open with JupyterLab's default handler instead.