You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Support for 'fill' attributes in 'scatter3d' #2352
FWIW that python example has a bug - looks like the fill is perhaps it's making a convex hull rather than just filling below the line? I haven't looked in enough detail to know whether this is a problem on the python or plotly.js side.
I don't think this is a Python specific issue. The surface fill here is being turned on by the scatter3d.surfaceaxis and scatter3d.surfacecolor properties.
The docs for surfaceaxis are
If "-1", the scatter points are not fill with a surface If 0,
1, 2, the scatter points are filled with a Delaunay surface
about the x, y, z respectively.
So I think what's happening here is that Delaunay triangularization doesn't take into account the order of the points and so when all of the points are on a plane you end up with a convex hull. And so it's not really meaningful to compare it with the line. So perhaps we should take down this documentation example.
I think what this Python example wants is the 3d fill option from the feature request.
FWIW, I have seen this bug in R as well. Posted an issue in documentation, then realized there is one already opened here, which is the place it should be. I just included the reproducible example in my thread.
So, it's 2024 and the feature is still not usable? Did anyone come up with a workaround? I see one solution to split one N-point polygon into N 4-points polygons, but that would mess with legend, create a lot of extra objects, etc.
Reacted by M--, Songyeon Lee, Steven Brown and Yahor Paromau
Activity
etpinard commentedon Feb 9, 2018
There's this thing: https://plot.ly/python/3d-filled-line-plots/ in the python docs.
Could be a nice feature for the R API? 😛
But yeah, why not, it would be nice to
fillattributes that behave more closely to what 2d axes support.alexcjohnson commentedon Feb 11, 2018
FWIW that python example has a bug - looks like the fill is perhaps it's making a convex hull rather than just filling below the line? I haven't looked in enough detail to know whether this is a problem on the python or plotly.js side.
jonmmease commentedon Oct 1, 2018
Reported again in plotly/plotly.py#1206
I don't think this is a Python specific issue. The surface fill here is being turned on by the
scatter3d.surfaceaxisandscatter3d.surfacecolorproperties.The docs for
surfaceaxisareSo I think what's happening here is that Delaunay triangularization doesn't take into account the order of the points and so when all of the points are on a plane you end up with a convex hull. And so it's not really meaningful to compare it with the line. So perhaps we should take down this documentation example.
I think what this Python example wants is the 3d
filloption from the feature request.MdoubleDash commentedon Apr 29, 2019
FWIW, I have seen this bug in R as well. Posted an issue in documentation, then realized there is one already opened here, which is the place it should be. I just included the reproducible example in my thread.
Zaesh commentedon Apr 28, 2020
Has anyone found a fix for that issue yet?
ak868308 commentedon Jan 29, 2021
Anybody found the solution for this using plotly-latest.js
positron96 commentedon Feb 26, 2024
So, it's 2024 and the feature is still not usable? Did anyone come up with a workaround? I see one solution to split one N-point polygon into N 4-points polygons, but that would mess with legend, create a lot of extra objects, etc.