How to use countplot() in plotly with VAEX data frame?

[bhargav-inthezone]

Some one please give me an alternate plotly code for this one :
sns.countplot(x='Census_ProcessorClass', hue='HasDetections',data=df_train)
plt.show()

both are int64

[nicolaskruchten]

This is basically px.histogram.

[bhargav-inthezone]

This is basically px.histogram.

df_train = vaex DataFrame
when I tried using this :

fig = px.histogram(df_train, x ='Census_ProcessorClass' , color = 'HasDetections', barmode = 'relative')
fig.show()

I am getting this Value error :
ValueError Traceback (most recent call last)
in
----> 1 fig = px.histogram(df_train, x ='Census_ProcessorClass' , color = 'HasDetections', barmode = 'relative')
2 fig.show()

/opt/conda/lib/python3.7/site-packages/plotly/express/_chart_types.py in histogram(data_frame, x, y, color, facet_row, facet_col, facet_col_wrap, facet_row_spacing, facet_col_spacing, hover_name, hover_data, animation_frame, animation_group, category_orders, labels, color_discrete_sequence, color_discrete_map, marginal, opacity, orientation, barmode, barnorm, histnorm, log_x, log_y, range_x, range_y, histfunc, cumulative, nbins, title, template, width, height)
454 histnorm=histnorm, histfunc=histfunc, cumulative=dict(enabled=cumulative),
455 ),
--> 456 layout_patch=dict(barmode=barmode, barnorm=barnorm),
457 )
458

/opt/conda/lib/python3.7/site-packages/plotly/express/_core.py in make_figure(args, constructor, trace_patch, layout_patch)
1859 apply_default_cascade(args)
1860
-> 1861 args = build_dataframe(args, constructor)
1862 if constructor in [go.Treemap, go.Sunburst] and args["path"] is not None:
1863 args = process_dataframe_hierarchy(args)

/opt/conda/lib/python3.7/site-packages/plotly/express/_core.py in build_dataframe(args, constructor)
1376
1377 df_output, wide_id_vars = process_args_into_dataframe(
-> 1378 args, wide_mode, var_name, value_name
1379 )
1380

/opt/conda/lib/python3.7/site-packages/plotly/express/_core.py in process_args_into_dataframe(args, wide_mode, var_name, value_name)
1181 if argument == "index":
1182 err_msg += "\n To use the index, pass it in directly as df.index."
-> 1183 raise ValueError(err_msg)
1184 elif length and len(df_input[argument]) != length:
1185 raise ValueError(

ValueError: Value of 'x' is not the name of a column in 'data_frame'. Expected one of [0] but received: Census_ProcessorClass

[nicolaskruchten]

Try converting your Vaex df to a Pandas one to see if that resolves things?

[bhargav-inthezone]

Try converting your Vaex df to a Pandas one to see if that resolves things?

Yeah Nic I am pretty sure it will resolve the issue but it will take a lot of time and memory to convert my data into pandas dataframe. I think my system may crash.

I am looking for more efficient ways. Is there any method to make Vaex dataframe acceptable by plotly.

[nicolaskruchten]

PX doesn't natively accept Vaex data frames at the moment, no. Part of the reason for that is that for plots like these histograms, it doesn't do Python-side aggregation: all the data is sent to the browser for aggregation, so there's a bit of an upper bound on the dataset size that px.histogram can handle anyway.

[nicolaskruchten]

See plotly/plotly.py#2649 for more details

[bhargav-inthezone]

See plotly/plotly.py#2649 for more details

Thanks will check this

[bhargav-inthezone]

See plotly/plotly.py#2649 for more details

Hey after lot of trail and errors, I think I found a better way. Check this code it worked

fig = px.histogram (x = df_train['Census_ProcessorClass'].tolist(), color= df_train['HasDetections'].tolist())
fig.show()

[bhargav-inthezone]

See plotly/plotly.py#2649 for more details

Hey after lot of trail and errors, I think I found a better way. Check this code it worked

fig = px.histogram (x = df_train['Census_ProcessorClass'].tolist(), color= df_train['HasDetections'].tolist())
fig.show()

I found a much better method:

df_train.select(df_train['Census_ProcessorClass'] ,'Census_ProcessorClass' != 'None' )
x_axis = df_train.evaluate(df_train['Census_ProcessorClass'], selection = True)
color_axis = df_train.evaluate(df_train['HasDetections'], selection = True)

%%time
fig = px.histogram (x = x_axis, color= color_axis, width = 300, height = 400)
fig.show()

CPU times: user 761 ms, sys: 33.1 ms, total: 794 ms
Wall time: 811 ms