Repository navigation
[FEATURE]: Allow dashed borders for scatter plot markers (marker.line.dash) #5499
Description
Description
I would like to request the ability to specify a dash style (e.g., "dash", "dot", "dashdot") for the border line of markers in scatter plots.
Currently, the marker.line object only supports properties like color and width. It does not support dash, unlike line traces or shapes.
Attempting to set marker=dict(line=dict(dash="dash")) currently raises a ValueError.
Why should this feature be added?
This feature is essential for creating clear visual distinctions in scatter plots without relying solely on color or marker symbols.
Use Case:
In data imputation or time-series analysis, it is common to visualize "missing" or "inferred" data points alongside original data.
- Original Data: Solid markers.
- Imputed/Inferred Data: Markers with a dashed outline (often hollow or lighter fill).
Current Limitations:
- Shapes: Using
layout.shapes(circles) allows for dashed borders, but shapes are defined in data coordinates. This causes circles to distort into ellipses when the x and y axes have different scales (common in time series). - Markers: Markers maintain their aspect ratio (perfect circles) and fixed pixel size regardless of zoom, which is desired, but they cannot (currently) have dashed borders.
Adding dash support to marker.line would bridge this gap, allowing for perfect, fixed-size circles with dashed outlines.
Mocks/Designs
Proposed API usage:
import plotly.graph_objects as go
fig = go.Figure(
go.Scatter(
mode="markers",
x=[1, 2, 3],
y=[1, 2, 3],
marker=dict(
size=20,
color="white",
line=dict(
color="red",
width=2,
dash="dash" # <--- Proposed Feature
)
)
)
)Example of what I am trying to achieve:
And this is as close as I can get with the current marker options (solid line):
import plotly.express as px
import pandas as pd
import numpy as np
dat = pd.DataFrame(
{
"time": [1, 2, 3, 4, 5, 6],
"all": [1, 7.5, 5.5, 4.8, 5.5, 7],
"missing": [False, False, True, True, False, False],
},
).assign(
value=lambda df: np.where(df["missing"], np.nan, df["all"]),
missing=lambda df: np.where(df["missing"], df["all"], np.nan),
)
(
px.line()
.add_scatter(
x=dat["time"],
y=dat["all"],
mode="lines",
name="all",
line=dict(shape="spline", smoothing=1.3, color="teal", width=3, dash="dash"),
)
.add_scatter(
x=dat["time"],
y=dat["value"],
mode="markers",
name="value",
marker=dict(color="teal", size=30),
)
.add_scatter(
x=dat["time"],
y=dat["missing"],
mode="markers",
name="missing",
marker=dict(
symbol="circle",
color="lightgrey",
size=30,
line=dict(
color="teal",
width=2,
# dash="dash", # <-- Proposed feature to be added here
),
),
)
.update_layout(
showlegend=False,
xaxis=dict(showgrid=False, showticklabels=False, showline=False, ticks=""),
yaxis=dict(showgrid=False, showticklabels=False, showline=False, ticks=""),
)
.show()
)
Notes
This seems to be a limitation in the underlying plotly.js library as well, so this might require changes upstream.
Activity
[-][FEATURE]:[/-][+][FEATURE]: Allow dashed borders for scatter plot markers (`marker.line.dash`)[/+]ndrezn commentedon Dec 10, 2025
We'll need to contribute this upstream in Plotly.js.
marker.line.dash) plotly.js#7667ndrezn commentedon Dec 10, 2025
Once added to Plotly.js, we can add a new ticket to expose it in Plotly.py
chrimaho commentedon Feb 3, 2026
@ndrezn, can you move this issue back to plotly/plotly.py? And re-open it. It looks like plotly/plotly.js#7673 has been approved and merged, and will soon be released. So this issue can be added to and implemented in the Ploytly.py repo.
camdecoster commentedon Feb 9, 2026
@chrimaho do you want to take a crack at adding this here as well?
chrimaho commentedon Feb 9, 2026
Yes! Absolutely 💯
Now that this issue is in the right repo, I'll fork it from here.
As soon as the new feature is released in plotly.js, I'll work on it, and raise a pr for plotly.py 👍
Thanks so much!
alexcjohnson commentedon Feb 9, 2026
Does this actually require plotly.py work to expose it? Won't it just be automatically exposed next time we rebuild?
chrimaho commentedon Feb 10, 2026
Hi @alexcjohnson, this is precisely what I will check and validate.
If no changes needed to plotly.py, then we can close this issue. If a change is needed, then I will raise a PR accordingly.
For now, I will wait until plotly.js releases the next update, then report back here with an update.
emilykl commentedon Aug 30, 2026
Addressed in plotly.js by plotly/plotly.js#7673 and available in plotly.py starting with v6.6.0.