Plotting into an existing axes replaces axes limits

[jl-wynen]

When plotting multiple data arrays into the same MPL Axes, the axes limits are replaced. E.g.,

import scipp as sc
import matplotlib.pyplot as plt
a = sc.DataArray(sc.ones(sizes={'x': 3}), coords={'x': sc.arange('x', 3)})
b = 10 * a

fig, ax = plt.subplots()
a.plot(ax=ax)
b.plot(ax=ax)

produces

that is, the points of a are not visible.

I would expect this to behave more like MPL where we can use ax.plot multiple times and get limits that encompass all data.

[nvaytet]

This is probably because it's difficult to keep everyone happy.

It comes from the fact that when doing b.plot(ax=ax), the figure created there is unaware of what was done to ax before.
In your case, you can either

fig, ax = plt.subplots()
a.plot(ax=ax)
b.plot(ax=ax)
ax.autoscale()

or use

import plopp as pp
pp.plot({'a': a, 'b': b})

[nvaytet]

@jl-wynen do you still want to modify the behaviour or are the suggested fixes above enough?

[jl-wynen]

In principle they are enough. I just find it surprising that it behaves differently from plain matplotlib.