如何在 Altair 条形图中突出显示具有不同 colors 的 2 个条形图？

Question

I would like to highlight two different bars in an Altair bar graph. However, each one has to have a different color, based on a different condition.

The basics:

import altair as alt
import pandas as pd

df = pd.DataFrame({"fruits": ["banana", "plum", "apple", "lime"], "values": [10, 6, 8, 5]})

plot = (
    alt.Chart(df)
    .mark_bar()
    .encode(
        x=alt.X("fruits", sort="y"),
        y="values"))

apple_plot = plot.encode(color=alt.condition(alt.datum.fruits == "apple", alt.value("red"), alt.value("navy")))

banana_plot = plot.encode(color=alt.condition(alt.datum.fruits == "banana", alt.value("yellow"), alt.value("navy")))

apple_plot | banana_plot

The result so far

What I tried

apple_banana_plot = plot.encode(
    color=alt.condition(
        alt.datum.fruits == "apple",
        alt.value("red"),
        alt.condition(
            alt.datum.fruits == "banana",
            alt.value("yellow"),
            alt.value("navy"))))

apple_banana_plot

The error I got:

TypeError: type object got multiple values for keyword argument 'condition'

...which means I cannot use a condition inside another.

The result I expect

Answer 1

You can do it like this: Basically, you define color with two parameters:

• domain - the list of values in the categorical column
• range - specific values of colors

df = pd.DataFrame({"fruits": ["banana", "plum", "apple", "lime"], "values": [10, 6, 8, 5]})

plot = (
    alt.Chart(df)
    .mark_bar()
    .encode(
        x=alt.X("fruits", sort="y"),
        y="values",
    color=alt.Color('fruits',
                     scale=alt.Scale(
                         domain=df.sort_values(['values'])['fruits'].tolist(),
                         range=['navy', 'navy', 'red', 'yellow']))))
plot