Getting 'TypeError: 'float' object cannot be interpreted as an integer' when using hvplot.box plot

hvPlot
#1

Hello,
I have a large dask.dataframe.core.DataFrame, with a column ‘AGE’, and dtype float:
AGE float64

When I want to create a box plot as below, I’m getting an error:
df.hvplot.box(y=‘AGE’)

~/anaconda/envs/eda-with-holoviz/lib/python3.7/site-packages/holoviews/plotting/bokeh/stats.py in _box_stats(self, vals)
157 vals = vals[is_finite(vals)]
158
–> 159 if len(vals):
160 q1, q2, q3 = (percentile(vals, q=q) for q in range(25, 100, 25))
161 iqr = q3 - q1

TypeError: ‘float’ object cannot be interpreted as an integer

Can you please help?

#2

Here is how you can reproduce the problem:

from dask import dataframe as dd

Create test dataframe

data = {‘Name’:[‘Tom’, ‘nick’, ‘krish’, ‘jack’], ‘Age’:[20, 21, 19, 18]}
df_test = pd.DataFrame(data)

Convert to dask dataframe

sd = dd.from_pandas(df_test, npartitions=1)
print(type(sd))
print(sd.head())

box plot

sd.hvplot.box(y = ‘Age’)

versions

$ pip list | grep viz
holoviz 0.11.6
pyviz-comms 0.7.6

#3

I got the same problem while executing the example code in the Box-Whisker Plots chapter (here)

flight_subset = flights[flights.carrier.isin([b'AA', b'US', b'OH'])] flight_subset.hvplot.box('depdelay', by='carrier', ylim=(-10, 70))

with holoviz 0.11.6

#4

Could you report your versions of HoloViews and hvPlot?