holoviz
.interactive dataframe repr: table columns overlap
ALL software version info
hvplot : 0.8.0
pandas : 1.4.3
holoviews : 1.14.9
bokeh : 2.4.3
Python version : 3.9.13
IPython version : 8.4.0
jupyter notebook : 6.4.12
jupyterlab : 3.4.3
OS : Darwin ( = MacOS 11.5.2)
Release : 20.6.0
Browser : Safari
Description of expected behavior and the observed behavior
The columns of the dataframe representation in both jupyter notebook and jupyter lab are too narrow, so numbers overlap for a particular example dataframe.
Complete, minimal, self-contained example code that reproduces the issue
import hvplot.pandas
from bokeh.sampledata import airports
df = airports.data
df1 = df.interactive()
df1 # show dataframe
Performing a dynamic calculation on the dataframe doesn't change that behavior
p = pnw.IntSlider(start=0, end=100)
df1[df1.index<p] # dataframe still squished
Stack traceback and/or browser JavaScript console output
No error message in jupyter notebook / lab. Ne error / warning in javascript console.
Screenshots or screencasts of the bug in action
Expected behavior
Update:
It seems like there is a width parameter for .interactive(). Unfortunately .interactive() doesn't seem to have a docstring, so it's far from obvious that there are parameters that you can give it, but I think I found it being used somewhere, that's why I thought about trying it.
With that width parameter the width of the dataframe table output can be controlled. the result is not perfect (because every column gets equal width), but it gets rid of the overlap.
I see there is a height parameter also. Would be nice if that would go in the docstring of .interactive().
@maximlt Seems to be solved now. So close it? An informative docstring was also added to .interactive, though height is not mentioned as a parameter, but width at least.