Plotly chart bars are still shown in stack with barmod defined to group

I want to display total and available bars side by side for each fruit.

Here is the data I'm using

import plotly.express as px

data_frame = {
    "fruits": ["orange", "apple", "banana", "cherry"],
    "total": [10, 20, 30, 25],
    "available": [7, 10, 27, 15],
}

And I used plotly.express as follow:

fig = px.bar(
    data_frame,
    x="fruits",
    y=["total", "available"],
    color="fruits",
    barmode="group", # Is ignored
)

fig.show()

However, bars are still shown in stack mode instead. What am I missing here?

Thanks in advance.

There might be a way to do this without changing your data from wide to long format, but I think it's more intuitive to melt your dataframe so that you have variable and value columns using:

df_melted = df.melt(id_vars='fruits', value_vars=['total','available'])

This gives you the following long format dataframe.

   fruits   variable  value
0  orange      total     10
1   apple      total     20
2  banana      total     30
3  cherry      total     25
4  orange  available      7
5   apple  available     10
6  banana  available     27
7  cherry  available     15

Then you can create the bar chart using:

fig = px.bar(df_melted, x='fruits', y='value', color='variable', barmode='group')

This will allow plotly to differentiate the between total and available bars within each fruit.