TableViewer

View table-like data files directly in Visual Studio Code.

TableViewer opens pickle, CSV/TSV, NumPy, HDF5, and Zarr files in a paginated table view. It is designed for quickly inspecting local data files without leaving the editor.

Features

• Opens .pkl, .pickle, .csv, .tsv, .tab, .npy, .h5, .hdf5, .hdf, and .zarr files.
• Detects comma-delimited and tab-delimited text, including tab-delimited .csv files.
• Supports extensionless pickle files without a wildcard custom editor selector: normal files open in VS Code first, and extensionless pickle files are redirected after a header check.
• Leaves unsupported files, dotfiles, source files, and images to VS Code's default editors.
• Handles HDF5 and Zarr files with multiple datasets by prompting for a data source.
• Provides pagination, configurable page size, multi-column sorting, pinned columns, stable wide-table horizontal scrolling, and search/filter operations.
• Uses lazy page loading for large NumPy arrays, generic HDF5 datasets, and Zarr arrays.

Requirements

This extension runs a local Python process to read data files.

Required Python packages depend on the file type:

• pandas: pickle DataFrame/Series and CSV/TSV files
• numpy: .npy files
• tables: pandas HDF5 stores
• h5py: generic HDF5 datasets
• zarr: Zarr arrays

You can configure the interpreter with the tableViewer.pythonPath setting or run Select Python Interpreter for TableViewer from the Command Palette. The Browse action opens a lightweight directory picker inside VS Code that shows folders plus files named python, python3, or python3.x, so extensionless executables such as /opt/conda/bin/python3 stay visible without showing unrelated files.

Usage

Open a supported file in VS Code. The file opens in TableViewer automatically for supported extensions.

You can also use the Explorer context menu command:

Open with TableViewer

For extensionless files, VS Code opens the file normally first. TableViewer only redirects to the viewer when the file header looks like pickle; non-pickle extensionless files stay in the normal editor.

Security

Pickle files can execute code when loaded. Only open pickle files from trusted sources.

TableViewer shows a confirmation prompt before loading .pkl and .pickle files.

Settings

{
  "tableViewer.pythonPath": "",
  "tableViewer.openExtensionlessByDefault": true
}

• tableViewer.pythonPath: Python interpreter path. Leave empty to auto-detect.
• tableViewer.openExtensionlessByDefault: For files without an extension, check whether the file looks like pickle before opening it in TableViewer.

Existing settings from older builds under pickleViewer.* are still read as a fallback when the new tableViewer.* setting is not explicitly set.

Notes

Large lazy-loaded arrays disable full-data sorting and filtering to avoid loading the entire file into memory.