5013 is in the sandbox, the functionality would be useful as df.inspect(),
popping up a gui grid view of the frame. In the future, it's possible to detect ipnb and render a js grid instead, if the coming ipython js functionality is adequate, to cover the remote kernel use-case (less common).
cc @alefnula
Comment From: ghost
cc @jtratner re https://groups.google.com/forum/#!topic/pydata/lgEmod8tGHs
Comment From: jtratner
@y-p - very cool
Comment From: jseabold
Can I vote for df.browse()? Having this feature in pandas would be great, even if it's incomplete. I assume just having it advertised would draw substantial interest. It's the one "big feature" I still miss. Unfortunately, my GUI skills are not up to snuff to make a serious go at this.
Comment From: TomAugspurger
https://github.com/ipython/ipython/blob/master/examples/widgets/Variable%20Inspector.ipynb Has an example that could serve as a template. I'll hack something together today.
What sort of features are you looking for?
I'm thinking options to color min / max values. Flagging NaNs. What else?
Comment From: jseabold
I'm personally more interested in something that works outside the notebook at the moment. It doesn't look like that one will?
As far as features, I'd just love to be able to do
df.browse()
And have a window pop-up that I can quickly scroll through up-down, left-right. Don't need anything else. Just a quick way to browse data will go a long way. The other linked example in the docs from #5013 is pretty close, though it looks like there are problems with setting data, etc. and it needs a bit of layout work. Advanced features don't need to be ready for something like that to be useful though.
Comment From: TomAugspurger
OK, I missed the part about it being outside the notebook. That's why I was asking about features, about what would differentiate it from the html repr. That one I linked to creates a draggable window inside the notebook.
Comment From: jreback
http://nbviewer.ipython.org/gist/rossant/9463955
Comment From: jorisvandenbossche
I think this is out of scope for core pandas, therefore closing (but it would certainly be nice to see better plugins for jupyter notebooks / jupyterlab)