MarcoGorelli
commented
2 days ago Wonderful, thanks @Trollgeir for trying out the prerelease, really appreciate it!
Indeed, here's reproducible example:
import plotly.express as px
import polars as pl
data = {
"Task": ["Research", "Design", "Implementation", "Testing", "Deployment"],
"Start": ["2024-01-01", "2024-02-01", "2024-03-01", "2024-04-15", "2024-05-01"],
"Finish": ["2024-01-31", "2024-02-28", "2024-04-14", "2024-04-30", "2024-05-15"],
"Resource": ["Team A", "Team B", "Team A", "Team C", "Team B"]
}
df = pl.DataFrame(data).with_columns(pl.col('Start', 'Finish').str.to_date())
# Create the timeline
fig = px.timeline(
df,
x_start="Start",
x_end="Finish",
y="Task",
color="Resource",
title="Project Timeline",
labels={"Resource": "Team Assigned"}
)
fig.update_yaxes(categoryorder="total ascending") # Order tasks by start date
fig.show()throws
Traceback (most recent call last):
File "/home/marcogorelli/scratch/.venv/lib/python3.12/site-packages/plotly/express/_core.py", line 2126, in process_dataframe_timeline
df = df.with_columns(
^^^^^^^^^^^^^^^^
File "/home/marcogorelli/scratch/.venv/lib/python3.12/site-packages/narwhals/dataframe.py", line 1402, in with_columns
return super().with_columns(*exprs, **named_exprs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/marcogorelli/scratch/.venv/lib/python3.12/site-packages/narwhals/dataframe.py", line 118, in with_columns
self._compliant_frame.with_columns(*exprs, **named_exprs),
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/marcogorelli/scratch/.venv/lib/python3.12/site-packages/narwhals/_polars/dataframe.py", line 89, in func
getattr(self._native_frame, attr)(*args, **kwargs)
File "/home/marcogorelli/scratch/.venv/lib/python3.12/site-packages/polars/dataframe/frame.py", line 9211, in with_columns
return self.lazy().with_columns(*exprs, **named_exprs).collect(_eager=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/marcogorelli/scratch/.venv/lib/python3.12/site-packages/polars/lazyframe/frame.py", line 2029, in collect
return wrap_df(ldf.collect(callback))
^^^^^^^^^^^^^^^^^^^^^
polars.exceptions.SchemaError: invalid series dtype: expected `String`, got `date`
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/marcogorelli/scratch/t.py", line 14, in <module>
fig = px.timeline(
^^^^^^^^^^^^
File "/home/marcogorelli/scratch/.venv/lib/python3.12/site-packages/plotly/express/_chart_types.py", line 432, in timeline
return make_figure(
^^^^^^^^^^^^
File "/home/marcogorelli/scratch/.venv/lib/python3.12/site-packages/plotly/express/_core.py", line 2482, in make_figure
args = process_dataframe_timeline(args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/marcogorelli/scratch/.venv/lib/python3.12/site-packages/plotly/express/_core.py", line 2131, in process_dataframe_timeline
raise TypeError(
TypeError: Both x_start and x_end must refer to data convertible to datetimes..str.to_datetime should only be called if Start / Finish aren't already of Datetime dtype
This only affects Polars plots, pandas ones are unaffected
If the user supplies a dataframe with correct dtypes (datetime), it will err as it's trying to cast them from string.
https://github.com/plotly/plotly.py/blob/master/packages/python/plotly/plotly/express/_core.py#L2125-L2129
It would be better if the casting is conditional, because right now I have to cast my datetime back to string, just for the function to cast it back to datetime.
plotly version 6.0.0rc1