Bokeh Server with 3 y axes doesn't plot,but does for 2 y axes.Why?

My objective is to plot live data on multiple y axes(different ranges). The following code worked with 2 y axes but when I added a 3rd one, the browser was just a blank.An earlier basic bokeh plot with 4 y axes worked fine but this server version is not working.I dont know what mistake I am doing,please guide me.

The following code plots 2 temporary points then a 3rd permanent point. Uses ‘patch’ and ‘stream’ from ColumnDataSource to achieve this.
Please download the attached files.

from import curdoc
from bokeh.models import ColumnDataSource
from bokeh.plotting import figure
from bokeh.models import LinearAxis, Range1d
import numpy as np
def plotcharts(source,x,a,b,c):
    fig=figure(y_range=(0, 500)), x=x,y=a,size=12,color='red')

    fig.extra_y_ranges = {"foo1": Range1d(start=0, end=50)},x=x, y=b, size=12, color="blue",
    fig.add_layout(LinearAxis(y_range_name="foo1"), 'left')

    # configure 3rd axis
    fig.extra_y_ranges = {"foo2": Range1d(start=0, end=25)}, x=x, y=c, color="magenta",
    fig.add_layout(LinearAxis(y_range_name="foo2"), 'right')

    return fig
#Just a and b work fine but when 3rd axis c was added it doesnt !
fig=plotcharts(source,'x','a','b','c') #c==3rd axis

def update_data():
    global count
    y1 = np.random.randint(low=0, high=500)
    y2 = np.random.randint(low=0, high=50)
    y3 = np.random.randint(low=0, high=25) #y3==3rd axis

    if count%3==0:
        new_data = {
            'x': [count/3],
            'a': [y1],
            'b': [y2],
            'c': [y3]  #c==3rd axis

        l = len(df['a'])
        new_data = {
            'a': [(l - 1,y1)],
            'b': [(l - 1, y2)],
            'c': [(l - 1, y3)]  #c==3rd axis


*no attached files here sorry.


I’ve answered on SO:




On Thursday, January 25, 2018 at 9:34:57 PM UTC+7, revoltman wrote:

*no attached files here sorry.

Yes, Thanks Eugene for your help.

Last time bigreddot(from stackoverflow) had pointed out this same mistake. I didnt learn :frowning:

Thanks again.

Property values that are dicts have always been too easy to misconfigure. We've gotten rid of most other instances of them, but I don't think we can in this case. But, we could probably add a convenience method to Figure for adding new ranges so that users don't have to much with the dictionary directly. I'd suggest making an issue on GitHub about this.




On Jan 25, 2018, at 10:22, revoltman <[email protected]> wrote:

Yes, Thanks Eugene for your help.
Last time bigreddot(from stackoverflow) had pointed out this same mistake. I didnt learn :frowning:
Thanks again.

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
For more options, visit