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
When using a datetime x axis and two y-axes single boxplots appear as vertical lines:
This also happens with a single y-axis and multiple plot types:
If you add a second boxplot on the same y-axis both render properly:
Code example for case 1: https://codepen.io/anon/pen/zeEKEo
FWIW what's happening is we don't know what width to give the boxes, because there are no other boxes on the same subplot to to give a distance to the next box, so they get a width of 1 - which for dates means 1 millisecond, hence on a scale of days it looks like zero width. You can fix this with an explicit width attribute (added recently in #3234, as @etpinard mentioned in #3508)
@etpinard are you thinking we can have boxes on overlaying subplots consider each other in width calculations? That seems like it would work for the first example, with boxes on each axis, and offhand I don't see any cases you'd want the existing behavior. Would you want to do this for stacked subplots as well? ie consider all boxes on the same position axis? That would be useful if @lemurey's
first example were converted to stacked subplots, though I can also imagine cases it might constitute a breaking change...
The second example (box + scatter) I don't really see any solution for except explicit width; there's really nothing to guide us in scaling the box.
Hi - this issue has been sitting for a while, so as part of our effort to tidy up our public repositories I'm going to close it. If it's still a concern, we'd be grateful if you could open a new issue (with a short reproducible example if appropriate) so that we can add it to our stack. Cheers - @gvwilson
Activity
etpinard commentedon Feb 6, 2019
Thanks for the report.
Very related to #3508
alexcjohnson commentedon Feb 6, 2019
FWIW what's happening is we don't know what width to give the boxes, because there are no other boxes on the same subplot to to give a distance to the next box, so they get a width of 1 - which for dates means 1 millisecond, hence on a scale of days it looks like zero width. You can fix this with an explicit width attribute (added recently in #3234, as @etpinard mentioned in #3508)
@etpinard are you thinking we can have boxes on overlaying subplots consider each other in width calculations? That seems like it would work for the first example, with boxes on each axis, and offhand I don't see any cases you'd want the existing behavior. Would you want to do this for stacked subplots as well? ie consider all boxes on the same position axis? That would be useful if @lemurey's
first example were converted to stacked subplots, though I can also imagine cases it might constitute a breaking change...
The second example (box + scatter) I don't really see any solution for except explicit width; there's really nothing to guide us in scaling the box.
lemurey commentedon Feb 6, 2019
@alexcjohnson Thanks for the note on width, I hadn't seen that, that will easily solve my immediate issue.
gvwilson commentedon Jun 11, 2024
Hi - this issue has been sitting for a while, so as part of our effort to tidy up our public repositories I'm going to close it. If it's still a concern, we'd be grateful if you could open a new issue (with a short reproducible example if appropriate) so that we can add it to our stack. Cheers - @gvwilson