I’m a long-time fan of Visidata. I initially planned to contribute directly upstream, but getting changes accepted can take time, while I wanted to use them myself and share sooner. That’s why I created a fork called Visidata NG (new generation). I’m developing it independently and proposing the most promising changes upstream in parallel.

At this point, it’s not published to pip/brew or other registries — it’s for enthusiasts who want to try new features early.

What I added

  1. Exit confirmation on the last stack level (q)
    When you’re at the last table in the stack and press q, Visidata NG now shows a confirmation dialog to prevent accidental exit and loss of unsaved changes. I used to lose my markup by mistake — not anymore.

  2. Quick “percent of total” via g+f
    Pressing g+f creates a new column based on the current one (if its type is int or float) and calculates the percent of the total. Frequency tables do show percentages by default (as a histogram), but there was no one-shot hotkey to create a dedicated column — now there is.

  3. Currency formatting on $ (column type currency)
    On $, a currency picker pops up, and the column is formatted according to the selected currency standard (prefix + value + suffix). Handy when you need quick stats and a neat Markdown table.

  4. Flexible date formats on @ (column type date)
    On @, you get a format picker with datetime, date, month, and year. This makes ad‑hoc exploration faster — no long expressions, just a couple of keystrokes.

Roadmap

Next, I’m considering going deeper into data storage/processing internals: vectorized operations could significantly speed up computations. For now, I focused on the practical tweaks I miss the most day-to-day.

How to try Visidata NG

  1. Clone the repository:
    git clone https://github.com/denisotree/visidata-ng
  2. Switch to the active development branch:
    cd visidata-ng
git checkout develop-ng
  3. Install in editable mode:
    pip install -e .
  4. The executable will be installed into the bin directory of your current Python environment.

Installation caveats

  • If you installed Visidata outside of pip (e.g., brew, apt, pacman, etc.), versions may conflict — you might need to remove the system-level package.
  • If Visidata was installed via pip, my fork will replace the original since it shares the same package metadata.

Feedback

Have ideas for improvements or found a bug? Message me (LinkedIn / Instagram / Telegram).