Legend colors missing

I have Vega3 installed and using the latest Jupyter notebook I am attempting to plot a bar chart with a legend. While colors are shown correctly in the chart, the corresponding colors are missing in the legend labels.

alt.Chart(df).mark_bar().encode(
    alt.X('subset:O', axis=alt.Axis(title='Subset')),
    alt.Y('sum(sha256):Q', axis=alt.Axis(title='Number of samples')),
    alt.Color('label:N', scale=alt.Scale(range=["#3333ff", "#ff3333"]), legend=alt.Legend(values=["Cat1", "Cat2"]))
)

[jakevdp]

Ah – so you’re setting the legend values to values that are not in your dataset, so there are no associated markers for those values.

If you want your legend to reflect labels that do not exist in your data, the best approach is to transform your data. You can either do this as a preprocessing step (i.e. modify the actual dataframe) or by using a transform within the chart specification. As an example of the latter approach, see this example figure where a column of zeros and ones is mapped to “Male” and “Female” within the chart.

[mozark24]

Here is a minimal dataframe which can be re-created to produce the problem.

df = pd.DataFrame({'label':[0,0,1,1],
                   'subset':['test','train','test','train'],
                   'sha256':[100000,300000,100000,300000]})