This is pure witchcraft! A single line of code makes my Pandas code 5x faster! I don't have to migrate to a different library or change anything else. The magic happens by patching Pandas with FireDucks, a multi-threaded, compiler-accelerated library with a fully compatible pandas API. The library makes... your Pandas code up to 48x faster! You have a couple of options to use FireDucks: 1. import fireducks.pandas as pd 2. Or run your code like this: `python -mfireducks.imhook main .py ` FireDucks is faster than Pandas and faster than Polars. Check the video attached to this post, and the link to the code below.show more

Here is a link to the code I show in the video: Here is FireDucks' GitHub Repository: This notebook compares FireDucks with Polars and Pandas: Finally, a few benchmarks:

If you prefer to watch the video on YouTube, here is the link:

This is the biggest productivity cheat code right now. Kiss reading documents goodbye. You can get an instant summary of any document with this tool.

One quick question: Is this speedup due only to multithreading? Also, have you tested Polars? What’s your opinion on it?

Your vey helpfull for so many peopels🔥

Why does pandas try to reinvent the wheel when SQL and @duckdb already do the job?

You forgot to mention its for Linux. Ouch, killed your post.

Hey there, you should try using pandas on on a gpu. CuDF is even crazier.

"Friendly reminder that even pandas need optimized bamboo-munching techniques to boost their performance!" 🐼🎋

Does it work on mac and windows?

What prevents pandas to adopt the optimizations?