plotly.express.bar does not work with pandas uint32 dtype when passing in a list of values for y

plotly.express.bar does not display properly when using pandas "uint32" dtype and passing a list for y


temp = pd.DataFrame({"my_category":list("ABCD"),"feature_1":[10, 20, 40, 30],"feature_2":[60, 40, 50, 20]})

for col in temp.columns:
    if temp[col].dtype == "int64":
        temp[col] = temp[col].astype("uint32")

px.bar(
    data_frame = temp,
    x = "my_category",
    y = ["feature_1","feature_2"],
)

using the "int64" dtype or plotting a single y works correctly

This last image is the expected behavior

plotly.version = 5.16.1
pandas.version = 2.0.2

Looks like this is the same root cause as #4291, with the same potential fix.

Pasting the relevant portion of my comment from #4291:

Looks like the bug traces to the _is_continuous() function in plotly/express/_core.py, which evaluates to False in the case of Rain_count, causing Plotly Express to decide that Rain_count is the categorical array and should be treated as the bar labels.

It seems the root cause is that dtype.kind for the Rain_count column is u, which is not one of "ifc".

Plausibly we should just change "in "ifc" to in "ifcu" — @MarcoGorelli do you see any issue with that? Looks like that does resolve the issue.

Numpy dtype.kind values: https://numpy.org/doc/stable/reference/generated/numpy.dtype.kind.html

It looks like this bug is also caused by uint32 columns not being treated as numeric.

fix px treating unsigned int columns as categorical

Navigation Menu

Repository navigation

plotly.express.bar does not work with pandas uint32 dtype when passing in a list of values for y #4344

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

plotly.express.bar does not work with pandas uint32 dtype when passing in a list of values for y #4344

Description

Activity

emilykl commented on Sep 1, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions