SooOOooo...just how FAST is this PYTHON library for Data Science?

Want a tutorial

yes please let's have a full polars tutorial to analyse a kaggle dataset

Kaggle competition tutorial

Yes Sir Nick, Kaggle competition, If it can be a live stream

Thanks

Using the right tool for the job: DuckDB.

Show us a vectorized operation (simd) operation. Unless you're going to compare that to polars there's really no point

Maybe do a better test 😂

Hey man is there any difference if we install polars from shell and directly on the jupyterlab

Milliseconds.

NO

Wow that's fast

Import polars as pd 🎉

i think bro doesn't know how fast is rust

Just use chunking with pandas and you're good to go

How to build a diffusion model?

Yes Sirrrrrr

We want Polars🎉 tutorial

Next library :-
Bears.
Then?
Grizzly bears.

What about pickle?

Kaggle 💪

Yes pls

But what's the catch?

why jupiter?

Yes we need

Let's see you run it

Yeah sure need a polar library tutorial ❤❤.

First tell us how to use black theme for jupiter lab

nice bro. i didn't leave a negative comment

Polars always crashes the kernel, unlike pandas

Have you created any end to end follow along tutorials? I'd like to learn by following along if possible. Thanks for the content!

would like to see a full video of how the big is difference for other methods

WRONG ADVICE!

If you are new to pandas stick with it, learn enough to be considered average, then move on to polars.

- don't spend time learning every new technology, you will end up learning nothing

- polars is fast and new, but Pandas is still the one with the most developed ecosystem, community, and job offers.

waiting for the tutorial

Polars tutorial abeg😂😂

Tips for the lazies like me: ask chatgpt to migrate your code snippets from pandas to Polars 😂

Have you tested how much memory is consumed? Sometimes performance is not everything, I can use all my computer resources to run an application but what is the cost? You end up having a super laggy computer as your memory is mostly consumed

Pyspark >>

polars also more consistent with the time per op

Do you know why?

Definitely

Indeed

polars uses lazy loading, so it isn't a fair comparison.

How fast it is? 5% of the c++ speed, what a fucking joke

Please do a tutorial on this for cnn-lstm hybrid model for image classification

theres pandas 2.0 coming up....check out the beta version

My only question is in what use case would you need something faster than pandas? I have been doing data science for a couple of years now and even the biggest datasets don't need that much time to load on pds

3rd party libraries? Fucking cringe

Is it better/faster than R ?

Then why not use polar pandas?

you lost me when you said rust