pandas-dev/pandas · error · ValueError

"{name}" is not a supported function

Error message

"{name}" is not a supported function

What it means

Raised by FuncNode.__init__ in pandas.core.computation.ops when a function call inside an eval/query expression names something not in MATHOPS (the union of _unary_math_ops like sin,cos,exp,log,... and _binary_math_ops arctan2). FuncNode wraps the call and binds it to the corresponding numpy function via getattr(np, name). The error is a ValueError.

Source

Thrown at pandas/core/computation/ops.py:561

class MathCall(Op):
    def __init__(self, func, args) -> None:
        super().__init__(func.name, args)
        self.func = func

    def __call__(self, env):
        # error: "Op" not callable
        operands = [op(env) for op in self.operands]  # type: ignore[operator]
        return self.func.func(*operands)

    def __repr__(self) -> str:
        operands = map(str, self.operands)
        return pprint_thing(f"{self.op}({','.join(operands)})")


class FuncNode:
    def __init__(self, name: str) -> None:
        if name not in MATHOPS:
            raise ValueError(f'"{name}" is not a supported function')
        self.name = name
        self.func = getattr(np, name)

    def __call__(self, *args) -> MathCall:
        return MathCall(self, args)

View on GitHub (pinned to 71959b8cb9)

Solutions

1. Use only whitelisted math functions: sin, cos, tan, exp, log, expm1, log1p, sqrt, sinh, cosh, tanh, arcsin, arccos, arctan, arccosh, arcsinh, arctanh, abs, log10, floor, ceil, arctan2.
2. For non-whitelisted functions, precompute the result into a column/variable and reference that in eval, or apply the function directly to the Series outside eval.
3. Use @local_func(arg) only if it resolves from scope - but note pure function calls inside eval still go through FuncNode, so prefer precomputing.

Example fix

# before
import pandas as pd
pd.eval('round(a, 2)')  # ValueError: "round" is not a supported function

# after (precompute)
import numpy as np
s = pd.Series([1.1, 2.6])
rounded = np.round(s, 2)   # apply directly
# or use a supported function:
pd.eval('floor(a)')        # 'floor' is whitelisted

from pandas.core.computation.ops import MATHOPS

def assert_supported_func(name: str) -> str:
    if name not in MATHOPS:
        raise ValueError(f'{name!r} not supported; whitelist: {MATHOPS}')
    return name

from pandas.core.computation.ops import MATHOPS

def is_supported_math_func(name: str) -> bool:
    return name in MATHOPS

try:
    pd.eval('foo(a)')
except ValueError as e:
    if 'not a supported function' in str(e):
        # precompute and pass as a variable
        a_computed = np.foo(a)
    raise

When it happens

Trigger: pd.eval('foo(a)') where 'foo' is not a whitelisted math function; df.query('isnan(a)'); pd.eval('len(a)'). Any function-call syntax in an eval string is checked against MATHOPS, and only those names resolve to numpy's implementations.

Common situations: Expecting arbitrary Python builtins (len, abs is allowed, round isn't) or numpy functions (isnan, isnan, vectorize) to be callable from eval. Also typoing a math function name (e.g. 'arcos' instead of 'arccos').

Related errors

• Only named functions are supported
• Invalid function call {node.func.id}
• Function "{res.name}" does not support keyword arguments
• keyword error in function call '{node.func.id}'
• The '@' prefix is only supported by the pandas parser

AI-assisted analysis of pandas-dev/pandas@71959b8cb9 (2026-08-07). Data as JSON: /api/errors/6abbfdce2d817b44. Report an issue: GitHub.