Duplicate factor error when stack vbar plots

Community Support
1

Hi, everyone! I’m a newbie using Bokeh. I had trouble using hover to show value and category together when I used vbar_stacks directly. Then I changed the dataframe(shown as below) and tried to loop vbar with CDSview for filtering category multiple times. But I got “Duplicate factor or subfactor error” in the output. Where did I do it wrong?

My code:

data = {‘prob’: prob,

‘prob_bottom’: prob_bottom,

‘prob_top’: prob_top,

‘preference_description’: preference_description,

‘preference_merchant’: preference_merchant}

source = ColumnDataSource(data=data)

hover = HoverTool(tooltips=[(‘description’,’@ preference_description’),

(‘prob’,’@prob’),

(‘merchant’,’@preference_merchant’)])

preference_description_unique = np.unique(preference_description).tolist()

p=figure(x_range= preference_description_unique,tools=[hover])

for i, merchant in enumerate(merchant_list):

view = CDSView(source=source, filters=[GroupFilter(column_name=‘preference_merchant’, group=merchant)])

p.vbar(x= ‘preference_description’, bottom = ‘prob_bottom’, top = ‘prob_top’, width=0.9, source = source, view = view, color = viridis(len(merchant_list))[i])

show(p)

···

My dataframe: Each ‘preference_description’ associates with 21 ‘preference_merchant’ and their probability values.

Many thanks!

2

Hi,

That error message is stating that you have not passed a list of unique factors for the x_range. I'm not sure why this is, given the code snippet you have below. It's not clear from your description what your individual factors look like. Does "preference_description' associates with 21 'preference_merchant' and their probability values." mean that the factors are 2-level, i.e. ["merchant", 0.1] or ["merchant", 0.4] ? If so, np.unique flattens its inputs, which is probably not what you intend. Otherwise, more information (ideally a reproducing script and data that can be run) is needed to say much more. I can only suggest printing preference_description_unique and verifying that it's actually what you expect, and that it has no duplicates.

Thanks,

Bryan

···

On Sep 29, 2017, at 22:29, Don Yang <[email protected]> wrote:

Hi, everyone! I'm a newbie using Bokeh. I had trouble using hover to show value and category together when I used vbar_stacks directly. Then I changed the dataframe(shown as below) and tried to loop vbar with CDSview for filtering category multiple times. But I got "Duplicate factor or subfactor error" in the output. Where did I do it wrong?

My code:

data = {'prob': prob,
            'prob_bottom': prob_bottom,
            'prob_top': prob_top,
            'preference_description': preference_description,
            'preference_merchant': preference_merchant}

source = ColumnDataSource(data=data)

hover = HoverTool(tooltips=[('description','@ preference_description'),
                                             ('prob','@prob'),
                                             ('merchant','@preference_merchant')])

preference_description_unique = np.unique(preference_description).tolist()
p=figure(x_range= preference_description_unique,tools=[hover])

for i, merchant in enumerate(merchant_list):
    view = CDSView(source=source, filters=[GroupFilter(column_name='preference_merchant', group=merchant)])
    p.vbar(x= 'preference_description', bottom = 'prob_bottom', top = 'prob_top', width=0.9, source = source, view = view, color = viridis(len(merchant_list))[i])

show(p)

-------------------------------
My dataframe: Each 'preference_description' associates with 21 'preference_merchant' and their probability values.
--------------------------------

Many thanks!

--
You received this message because you are subscribed to the Google Groups "Bokeh Discussion - Public" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/f69a984e-6c07-4b7a-80ea-8067340fdbd6%40continuum.io\.
For more options, visit https://groups.google.com/a/continuum.io/d/optout\.
<Snapshot of dataframe.png>

3

Hi Bryan,

Many thanks! I checked preference_description_unique but there is no any duplicate. The factor is 2-level (preference_description and preference_merchant) but I only passed preference_description_unique to x_range. My goal is to make stacked vbar with tooltips, showing preference_merchant and its probability. I attach data and code below. Thanks!

data.txt (57.1 KB)

···

from bokeh.io import show, output_file

from bokeh.models import ColumnDataSource, HoverTool, CDSView, GroupFilter

from bokeh.plotting import figure

from bokeh.palettes import viridis

output_file(‘./images/Preference.html’)

data = {‘prob’: prob,

‘prob_bottom’: prob_bottom,

‘prob_top’: prob_top,

‘preference_description’: preference_description,

‘preference_merchant’: preference_merchant}

source = ColumnDataSource(data=data)

hover = HoverTool(tooltips=[(‘description’,‘@preference_description’),

(‘prob’,‘@prob’),

(‘merchant’,‘@preference_merchant’)])

preference_description_unique = np.unique(preference_description).tolist()

p=figure(x_range=preference_description_unique,tools=[hover])

merchant_list = np.unique(preference_merchant).tolist()

for i, merchant in enumerate(merchant_list):

view = CDSView(source=source, filters=[GroupFilter(column_name=‘preference_merchant’, group=merchant)])

p.vbar(x= ‘preference_description’, bottom = ‘prob_bottom’, top = ‘prob_top’, width=0.9, source = source, view = view, color = viridis(len(merchant_list))[i])

show(p)

Best,

Don

在 2017年9月30日星期六 UTC-4上午10:28:22，Bryan Van de ven写道：

Hi,

That error message is stating that you have not passed a list of unique factors for the x_range. I’m not sure why this is, given the code snippet you have below. It’s not clear from your description what your individual factors look like. Does “preference_description’ associates with 21 ‘preference_merchant’ and their probability values.” mean that the factors are 2-level, i.e. [“merchant”, 0.1] or [“merchant”, 0.4] ? If so, np.unique flattens its inputs, which is probably not what you intend. Otherwise, more information (ideally a reproducing script and data that can be run) is needed to say much more. I can only suggest printing preference_description_unique and verifying that it’s actually what you expect, and that it has no duplicates.

Thanks,

Bryan

On Sep 29, 2017, at 22:29, Don Yang [email protected] wrote:

Hi, everyone! I’m a newbie using Bokeh. I had trouble using hover to show value and category together when I used vbar_stacks directly. Then I changed the dataframe(shown as below) and tried to loop vbar with CDSview for filtering category multiple times. But I got “Duplicate factor or subfactor error” in the output. Where did I do it wrong?

My code:

data = {‘prob’: prob,

        'prob_bottom': prob_bottom,

        'prob_top': prob_top,

        'preference_description': preference_description,

        'preference_merchant': preference_merchant}

source = ColumnDataSource(data=data)

hover = HoverTool(tooltips=[(‘description’,‘@ preference_description’),

                                         ('prob','@prob'),

                                         ('merchant','@preference_merchant')])

preference_description_unique = np.unique(preference_description).tolist()

p=figure(x_range= preference_description_unique,tools=[hover])

for i, merchant in enumerate(merchant_list):

view = CDSView(source=source, filters=[GroupFilter(column_name='preference_merchant', group=merchant)])

p.vbar(x= 'preference_description', bottom = 'prob_bottom', top = 'prob_top', width=0.9, source = source, view = view, color = viridis(len(merchant_list))[i])

show(p)

My dataframe: Each ‘preference_description’ associates with 21 ‘preference_merchant’ and their probability values.

Many thanks!

–
You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/f69a984e-6c07-4b7a-80ea-8067340fdbd6%40continuum.io.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.