Time Series xaxis ticks in 3.2.0

[sursu]

I have the following dataframe:

DD = pd.DataFrame({'Intervals':np.random.randint(30, size=1000)})
DD['Date'] = pd.Timestamp.today() + pd.to_timedelta(DD['Intervals'],'h').cumsum()
DD['Duration'] = np.random.normal(size=len(DD))

I try to make a time series plot as shown here:

trace = go.Scatter(x=DD['Date'], y=DD['Duration'],
                   mode='lines+markers', 
                   marker=dict(size = 4),
                   line=dict(width = 0.8))
layout = go.Layout(yaxis=dict(title='Seconds'),
                   height=700
                  )
fig = go.Figure([trace], layout=layout)
iplot(fig, show_link=False)

Output:

Notice how the xaxis ticks are not in a datetime format.

Specifying the x as x=DD['Date'].dt.date resolves the issue.
But this solution does not work if I use go.FigureWidget().
And, it is wrong, as it can be seen if I were to use 's' instead of 'h' as the unit in pd.to_timedelta().

---

I use Jupyter Lab. The same happens in Jupyter Notebooks.
This issue did not exist previously (e.g. in 3.1.2).

[jonmmease]

Hi @sursu , I think you're running into #1160, which was fixed in 3.2.1. Can you try updating and see if that takes care of it?

[sursu]

Yes, that solved it 👍