`px.pie` raises `ColumnNotFoundError` when `names` and `category_order` is passed at the same time on plotly 6.0.0

[schancksb]

Hey there, I've been having issues with the latest release, starting from v6.0.0, any pie plot made with plotly express raises a polars.exceptions.ColumnNotFoundError when passing the category_orders and names arguments at the same time.

A quick look at the traceback reveals a hardcoded column inside process_dataframe_pie function, which might be the source of the issue. I haven't tested this extensively and I don't know if it is by design or a bug.

In case anyone find themselves with this issue, for now I've found the following workarounds:

• Pin plotly to v5.24.1
• Sort the dataframe columns before passing it to px.pie and disable internal sorting with fig.update_traces(sort=False)
• Use go.Pie directly

Error traceback

2025-01-29 17:58:58.746 Uncaught app execution
Traceback (most recent call last):
  File "/usr/local/lib/python3.10/site-packages/narwhals/_polars/dataframe.py", line 109, in func
    getattr(self._native_frame, attr)(*args, **kwargs)
  File "/usr/local/lib/python3.10/site-packages/polars/dataframe/frame.py", line 9586, in with_columns
    return self.lazy().with_columns(*exprs, **named_exprs).collect(_eager=True)
  File "/usr/local/lib/python3.10/site-packages/polars/lazyframe/frame.py", line 2056, in collect
    return wrap_df(ldf.collect(callback))
polars.exceptions.ColumnNotFoundError: b

Resolved plan until failure:

        ---> FAILED HERE RESOLVING 'with_columns' <---
DF ["status", "count"]; PROJECT */2 COLUMNS

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/app/components/statistics/home_graphs.py", line 121, in object_by_status_pie_plot
    pie_plot = px.pie(
  File "/usr/local/lib/python3.10/site-packages/plotly/express/_chart_types.py", line 1701, in pie
    return make_figure(
  File "/usr/local/lib/python3.10/site-packages/plotly/express/_core.py", line 2481, in make_figure
    args, trace_patch = process_dataframe_pie(args, trace_patch)
  File "/usr/local/lib/python3.10/site-packages/plotly/express/_core.py", line 2165, in process_dataframe_pie
    df.with_columns(
  File "/usr/local/lib/python3.10/site-packages/narwhals/dataframe.py", line 1764, in with_columns
    return super().with_columns(*exprs, **named_exprs)
  File "/usr/local/lib/python3.10/site-packages/narwhals/dataframe.py", line 138, in with_columns
    self._compliant_frame.with_columns(*compliant_exprs, **compliant_named_exprs),
  File "/usr/local/lib/python3.10/site-packages/narwhals/_polars/dataframe.py", line 113, in func
    raise ColumnNotFoundError(msg) from e
narwhals.exceptions.ColumnNotFoundError: b

Resolved plan until failure:

        ---> FAILED HERE RESOLVING 'with_columns' <---
DF ["status", "count"]; PROJECT */2 COLUMNS

Hint: Did you mean one of these columns: ['status', 'count']?

Minimal example to reproduce this issue

import plotly.express as px
import polars as pl

df = pl.DataFrame({
    "status": ["On Route", "Pending", "Waiting Result", "Delivered"],
    "count": [28, 10, 73, 8]
})

custom_order = ["Pending", "Waiting Result", "On Route", "Delivered",]

fig = px.pie(
    data_frame=df,
    values="count",
    names="status",
    category_orders={"status": custom_order}, # raises polars.exceptions.ColumnNotFoundError
)

fig.show()

Potential Source
Hardcoded "b" column name on line 2166 of packages/python/plotly/plotly/express/_core.py

def process_dataframe_pie(args, trace_patch):
    names = args.get("names")
    if names is None:
        return args, trace_patch
    order_in = args["category_orders"].get(names, {}).copy()
    if not order_in:
        return args, trace_patch
    df: nw.DataFrame = args["data_frame"]
    trace_patch["sort"] = False
    trace_patch["direction"] = "clockwise"
    uniques = df.get_column(names).unique(maintain_order=True).to_list()
    order = [x for x in OrderedDict.fromkeys(list(order_in) + uniques) if x in uniques]

    # Sort args['data_frame'] by column 'b' according to order `order`.
    token = nw.generate_temporary_column_name(8, df.columns)
    args["data_frame"] = (
        df.with_columns(
            nw.col("b") # potential issue
            .replace_strict(order, range(len(order)), return_dtype=nw.UInt32)
            .alias(token)
        )
        .sort(token)
        .drop(token)
    )
    return args, trace_patch

[MarcoGorelli]

Thanks @schancksb for the report!

This might be due to 499e2fa - @FBruzzesi fancy taking a look? Should it be names instead of nw.col('b')?

EDIT in fact it looks like that commit was originally from https://github.com/FBruzzesi/plotly.py/pull/5/files , so this is my bad here 😳 Will make a fix + test first thing in the morning

[schancksb]

Wow that was fast, thanks for the fix 😃

[matthias-roussel-soyhuce]

Is there a workaround until this is published ? Besides ordering the original dataframe ?