Operator ">" not supported for types X and "Literal[0]"

[schorlton-bugseq]

Describe the bug

I have:
assert df.at["a", "b"] > 0

and am getting:

error: Operator ">" not supported for types "Scalar" and "Literal[0]"
    Operator ">" not supported for types "str" and "Literal[0]"
    Operator ">" not supported for types "bytes" and "Literal[0]"
    Operator ">" not supported for types "date" and "Literal[0]"
    Operator ">" not supported for types "datetime" and "Literal[0]"
    Operator ">" not supported for types "timedelta" and "Literal[0]"
    Operator ">" not supported for types "Timestamp" and "Literal[0]"
    Operator ">" not supported for types "Timedelta" and "Literal[0]"
    Operator ">" not supported for types "complex" and "Literal[0]" (reportOperatorIssue)

I would have thought this would be fine if .at returns an int or float (and this code runs fine) but maybe I'm missing something?

Please complete the following information:

• pandas-stubs==2.3.2.250926
• pyright v1.1.402

Thanks for your help!!

[cmp0xff]

I have made a runnable script:

import pandas as pd

df = pd.DataFrame({"b": 2}, index=["a"])

assert df.at["a", "b"] > 0  # pyright gives reportOperatorIssue; ruff complains https://docs.astral.sh/ruff/rules/pandas-use-of-dot-at/

[Dr-Irv]

The main issue here is that the type of df.at["a", "b"] could not only be a number, but a string, a datetime, etc. You can't compare a string to a number.

From a static typing perspective, the type of .at is ambiguous, so the only way to handle this is to tell the type checker the type that you are expecting for the expression df.at["a", "b"], e.g., cast(int, df.at["a", "b"]) > 0

Closing as there is nothing we can do here.