pandas-dev/pandas · error · TypeError

Array with ndim > 2 is not supported.

Error message

Array with ndim > 2 is not supported.

What it means

The rank() algorithm supports only 1-D and 2-D arrays. For 1-D input it calls rank_1d; for 2-D it calls rank_2d with an axis parameter. Any array with three or more dimensions falls through to this TypeError. This is a fundamental limitation — the ranking logic has no implementation for higher-dimensional data.

Solutions

1. Reshape the array to 1-D or 2-D before ranking: arr.reshape(-1) or arr.reshape(arr.shape[0], -1).
2. Apply rank() to individual Series or DataFrame columns rather than the raw high-dimensional array.
3. If you have panel data, use a MultiIndex DataFrame (2-D) instead of a 3-D numpy array.

Example fix

# before
arr_3d = np.random.randn(3, 4, 5)
pd.DataFrame(arr_3d).rank()

# after — reshape to 2-D first
arr_2d = arr_3d.reshape(arr_3d.shape[0], -1)
pd.DataFrame(arr_2d).rank()

def safe_rank(arr):
    arr = np.asarray(arr)
    if arr.ndim > 2:
        raise ValueError(f"rank supports max 2-D arrays, got ndim={arr.ndim}")
    return pd.DataFrame(arr).rank()

def is_rankable_ndim(arr) -> bool:
    return np.asarray(arr).ndim <= 2

try:
    result = pd.DataFrame(arr).rank()
except TypeError as e:
    if "ndim > 2" in str(e):
        arr = np.asarray(arr).reshape(-1)
        result = pd.Series(arr).rank()
    else:
        raise

When it happens

Trigger: Calling pd.Series.rank() or DataFrame.rank() on data that has been reshaped to 3+ dimensions. Passing a 3-D numpy array to pandas.core.algorithms.rank directly. Attempting to rank a panel-like or multi-dimensional numpy array without first flattening or selecting a 2-D slice.

Common situations: Reshaping data with np.reshape or xarray-to-numpy conversion that produces 3-D arrays, then passing them to pandas ranking functions. Working with image data or tensor representations. Building multi-index DataFrames from higher-dimensional arrays and calling rank before reducing dimensions.

Related errors

• bins argument only works with numeric data.
• can only perform ops with 1-d structures
• Column length mismatch
• 'data' must have a single column, not
• requires a Series, Index, ExtensionArray, np.ndarray or…

AI-assisted analysis of pandas-dev/pandas@3b7651241d (2026-08-11). Data as JSON: /api/errors/d8af9961b9723792. Report an issue: GitHub.

Appendix: source

Thrown at pandas/core/algorithms.py:1254

            ties_method=method,
            ascending=ascending,
            na_option=na_option,
            pct=pct,
            mask=mask,
        )
    elif values.ndim == 2:
        assert mask is None
        ranks = algos.rank_2d(
            values,
            axis=axis,
            is_datetimelike=is_datetimelike,
            ties_method=method,
            ascending=ascending,
            na_option=na_option,
            pct=pct,
        )
    else:
        raise TypeError("Array with ndim > 2 is not supported.")

    return ranks


def is_monotonic(values: ArrayLike) -> tuple[bool, bool, bool]:
    """
    Determine whether values are monotonic increasing/decreasing.

    Parameters
    ----------
    values : np.ndarray or ExtensionArray

    Returns
    -------
    tuple[bool, bool, bool]
        (is_monotonic_increasing, is_monotonic_decreasing, is_strict_monotonic)

    Raises

View on GitHub (pinned to 3b7651241d)