What causes "Can't determine version for {module.__name__}" in pandas-dev/pandas?

Reinstall the offending dependency from a clean wheel: `pip install --force-reinstall `.

pandas-dev/pandas · error · ImportError

Can't determine version for

Error message

Can't determine version for {module.__name__}

What it means

Raised by `get_version` in pandas.compat._optional when an imported module has no `__version__` attribute. pandas requires version metadata to enforce minimum-version checks for optional dependencies; without it, the safety check cannot run. The message names the offending module so the user can identify which dependency is misbehaving.

Solutions

Reinstall the offending dependency from a clean wheel: `pip install --force-reinstall <name>`.
Verify in a fresh interpreter: `python -c 'import <mod>; print(getattr(<mod>, "__version__", None))'`.
If the dependency genuinely lacks `__version__`, raise an issue upstream or pin a version that has it.
If you control the dependency, set `__version__` in its `__init__.py`.

Defensive patterns

Strategy: try-catch

Validate before calling

import importlib
m = importlib.import_module(name)
assert hasattr(m, '__version__'), f'{name} missing __version__'

Try / catch

try:
    from pandas.compat._optional import get_version
    ver = get_version(module)
except ImportError:
    ver = 'unknown'

Prevention

Reinstall dependencies from clean wheels in editable-install trouble.
Verify __version__ exists for any custom dependency before relying on version checks.

When it happens

Trigger: Importing an optional dependency whose package did not set `__version__` (some pure-Python packages or namespace packages); pre-release or local editable installs missing version metadata; a stub/shim module shadowing the real package.

Common situations: Editable installs (`pip install -e .`) of a dependency that don't bake in version; some conda environments shipping a module without `__version__`; namespace package collisions.

Related errors

AI-assisted analysis of pandas-dev/pandas@3b7651241d (2026-08-11). Data as JSON: /api/errors/f49819a2366e8b3d. Report an issue: GitHub.

Appendix: source

Thrown at pandas/compat/_optional.py:80

# these two names are different.

INSTALL_MAPPING = {
    "bs4": "beautifulsoup4",
    "bottleneck": "Bottleneck",
    "jinja2": "Jinja2",
    "lxml.etree": "lxml",
    "odf": "odfpy",
    "python_calamine": "python-calamine",
    "sqlalchemy": "SQLAlchemy",
    "tables": "pytables",
}


def get_version(module: types.ModuleType) -> str:
    version = getattr(module, "__version__", None)

    if version is None:
        raise ImportError(f"Can't determine version for {module.__name__}")
    if module.__name__ == "psycopg2":
        # psycopg2 appends " (dt dec pq3 ext lo64)" to its version
        version = version.split()[0]
    return version


@overload
def import_optional_dependency(
    name: str,
    extra: str = ...,
    min_version: str | None = ...,
    *,
    errors: Literal["raise"] = ...,
) -> types.ModuleType: ...


@overload
def import_optional_dependency(

View on GitHub (pinned to 3b7651241d)