Sankey node x and y positions could not be equal to 0 or np.nan

[banderlog]

1. We need to define both x and y or nothing will work Sankey node x and y positions require both variables set to reflect in chart. #1732
2. We cannot maintain the original position of a node (in R one could just add NA to that vector position)
3. x and/or y should not be equal exactly to 0

Consider example below

# example from <https://plotly.com/python/sankey-diagram/>
import plotly.graph_objects as go

fig = go.Figure(data=[go.Sankey(
    arrangement = "perpendicular",
    node = dict(
      pad = 15,
      thickness = 20,
      line = dict(color = "black", width = 0.5),
      label = ["A1", "A2", "B1", "B2", "C1", "C2"],
      color = "blue",
    ),
    link = dict(
      source = [0, 1, 0, 2, 3, 3], # indices correspond to labels, eg A1, A2, A1, B1, ...
      target = [2, 3, 3, 4, 4, 5],
      value = [8, 4, 2, 8, 4, 2]
  ))])

fig.update_layout(title_text="Basic Sankey Diagram", font_size=10)
fig.show()

Now we want to put A2, B2, C2 before A1,B1,C1 on Y axis:

fig.data[0].node.x = [0., 0., 0.5, 0.5, 1, 1]
fig.data[0].node.y = [1, 0.5, 1, 0.5, 1, 0.5]
fig.show()

We are able to see that A1 and A2 did not change their places. That because they have 0 in x coordinates.
It is easy to fix:

fig.data[0].node.x = [1e-09, 1e-09, 0.5, 0.5, 1, 1]

We are not reloading Ys, just fixing Xs is enough.
Probably plotly.js problem on visualization, but plotly.python problem for letting this.

If we will do something like fig.data[0].node.x = [np.nan, 1e-09, 0.5, 0.5, 1, 1], A1 will behave in the same way as if it would be a 0.