Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Exception raised with empty dataframe in v6.0.0rc0 #4919

Closed
KoolADE85 opened this issue Nov 29, 2024 · 3 comments · Fixed by #4922
Closed

Exception raised with empty dataframe in v6.0.0rc0 #4919

KoolADE85 opened this issue Nov 29, 2024 · 3 comments · Fixed by #4922
Assignees
@emilykl
Labels
bug something broken P1 needed for current cycle

Comments

@KoolADE85
Copy link

Our App Studio test suite caught an issue with this code in the latest RC of plotly.py:

df = pd.DataFrame({
  'x': [],
  'y': [],
})
fig = px.bar(df, x="x", y="y")

In v5.24.1, this produces an empty graph as expected
In v6.0.0rc0, this raises an Exception: ValueError: zero-size array to reduction operation maximum which has no identity

This is a simple case, but a more realistic use case would be when a dataframe is filtered down to nothing due to user input.
@MarcoGorelli
Copy link
Contributor

Thanks @KoolADE85 for reporting, and for trying out the prerelease!

I've tried this out but can't reproduce the error:

Image

Any chance you could share some more details please? Could you show the output of

import narwhals
narwhals.show_versions()

please?

@KoolADE85
Copy link
Author

My apologies on the poor example code... must have been my EOD Friday brain 😅

Here is code that DOES reproduce the error in a notebook. It's worth noting that the error occurs when fig is output (last line), and not when it is created.

df = pd.DataFrame({
  'x': ['a', 'b', 'c'],
  'y': [1 ,2, 3],
})
dff = df.query('x == "d"')
fig = px.bar(dff, x='x', y='y')
fig

Image

And here is the same code as a Dash app, where the graph would be JSON-encoded rather than output directly. Here, the graph displays until you select d from the dropdown list. The error message comes from the callback (i.e. it's server-side) and is identical to the error produced in the notebook example.

import pandas as pd
import plotly.express as px
from dash import Dash, Input, Output, callback, dcc, html

df = pd.DataFrame({
  'x': ['a', 'b', 'c'],
  'y': [1 ,2, 3],
})

app = Dash()
app.layout = [
  dcc.Dropdown(["a", "b", "c", "d"], value="a", id="inpt"),
  dcc.Graph(figure=px.bar(df, x='x', y='y'), id="outpt"),
]

@callback(
  Output("outpt", "figure"),
  Input("inpt", "value"),
  prevent_initial_call=True,
)
def cb(dropdown_selection):
  dff = df.query(f'x == "{dropdown_selection}"')
  return px.bar(dff, x='x', y='y')

server = app.server
if __name__ == "__main__":
  server.run(debug=True)

@KoolADE85
Copy link
Author

Output from narwhals.show_versions():

System:
    python: 3.12.7 (main, Nov 10 2011, 15:00:00) [GCC 14.2.0]
executable: ...
   machine: Linux-6.11.8-200.fc40.x86_64-x86_64-with-glibc2.40

Python dependencies:
     narwhals: 1.15.1
       pandas: 2.2.2
       polars: 1.14.0
         cudf: 
        modin: 0.32.0
      pyarrow: 17.0.0
        numpy: 1.26.4

@gvwilson gvwilson added bug something broken P1 needed for current cycle labels Dec 2, 2024
@emilykl emilykl mentioned this issue Dec 2, 2024
Merged
@emilykl emilykl closed this as completed in 43ac339 Dec 3, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@emilykl emilykl

Labels
bug something broken P1 needed for current cycle
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

fix: Skip base64 conversion for empty arrays (fixes #4919) plotly/plotly.py
4 participants
@gvwilson @emilykl @KoolADE85 @MarcoGorelli