Thank you! I'd suggest using dates for more than just the x-axis (currently, the three subplots are testing the same thing). A common pattern is something like:

• dates for x-axis
• dates for y-axis
• dates for both x- and y-axis

In this way, one can be a bit more certain that it works.

Hi!
I have changed my test as you requested. Please see the updated plot below. Thanks!

I think that an explicit value should be passed forbaseline here. the default value is the (arbitrary) zero mark for our date system (1980-01-01), and that is creeping into these plots.

That also tests that the baseline behavior acts as we expect with unitful data.

Hi! Thank you for your reply. I tried to use baseline but I found there is no where else using this baseline value in the test_datetime file. Is it necessary to use this value in this test?

baseline is a parameter specifically tostairs. thebottom argument tobar andstem is the same idea (just a different name).

(stackplot also had discussion aboutbaseline, though that spun out into#27129, because that method only supports specific values)

Thank you for the clarification! I have added the baseline component into my test in the latest commits. And you can see that the figure is now limited at 1980-1-1

Thank you for your suggestion! I have just committed your suggestions!

Thanks@ifEricReturnTrue! Congratulations on your first PR to Matplotlib 🎉 We hope to hear from you again.