When summing two charts with identically ordered axis, the order is lost.

Hi, I’m rather new to the library and this is probably derived from my lack of full understanding of it, but since this behavior is a little bit surprising I wanted to ask if it is expected.

I have two dataframes:

import altair as alt
import pandas as pd

laureates = pd.DataFrame({
    'name': ['B', 'D', 'C', 'A'],
    'born': pd.to_datetime(['1900-01-01', '1905-01-01', '1930-01-01', '1960-01-01']),
    'died': pd.to_datetime(['1950-01-01', '1965-01-01', '1970-01-01', '2015-01-01'])
})

prizes = pd.DataFrame({
    'name': ['B', 'B', 'D', 'C', 'A'],
    'date': pd.to_datetime(['1930-01-01', '1935-05-01', '1936-01-01', '1965-01-01', '2000-01-01'])
})

And I did the following two charts:

dots = alt.Chart(prizes.merge(laureates, on='name'))\
    .mark_point(fill='black')\
    .encode(y='date:T', x=alt.X("name:N", sort=["born:T"]))

lines = alt.Chart(laureates)\
    .mark_rule()\
    .encode(y="born:T", y2="died:T", x=alt.X("name:N", sort=["born:T"]))

Independently, both graphs render more or less as I intended:

But when I sum them with lines + dots the ordering in the x-axis reverts to alphabetical:

Is this expected? If so, how do I keep the original ordering? Or how do I reorder the axis after joining the charts?

[jakevdp]

This is a bug in Vega-Lite; see #820 and vega/vega-lite#3710. Closing as a duplicate.

[jeffbaumes]

As @jakevdp suggests in #820, you can update one of your alt.X specifications to turn off the second axis:

alt.X("name:N", sort=["born:T"], axis=alt.Axis(labels=False, ticks=False))