What causes "unknown method '{method}'" in jax-ml/jax?

Use one of: 'average', 'min', 'max', 'dense', 'ordinal' (exact lowercase).

jax-ml/jax · error · ValueError

unknown method '{method}'

Error message

unknown method '{method}'

What it means

jax.scipy.stats.rankdata only supports the tie-handling methods 'average', 'min', 'max', 'dense', and 'ordinal'; any other method string raises ValueError.

Source

Thrown at jax/_src/scipy/stats/_core.py:197

  if nan_policy not in ["propagate", "omit", "raise"]:
    raise ValueError(
      f"Illegal nan_policy value {nan_policy!r}; expected one of "
      "{'propagate', 'omit', 'raise'}"
    )
  if nan_policy == "omit":
    raise NotImplementedError(
      f"Logic for `nan_policy` of {nan_policy} is not implemented"
    )
  if nan_policy == "raise":
    raise NotImplementedError(
      "In order to best JIT compile `rankdata`, we cannot know whether `x` "
      "contains nans. Please check if nans exist in `x` outside of the "
      "`rankdata` function."
    )

  if method not in ("average", "min", "max", "dense", "ordinal"):
    raise ValueError(f"unknown method '{method}'")

  if axis is not None:
    return jnp.apply_along_axis(rankdata, axis, a, method)

  a = jnp.ravel(a)
  out_dtype = dtypes.default_float_dtype()

  def _rankdata(a: Array) -> Array:
    arr, sorter = lax.sort_key_val(a, jnp.arange(a.size))
    inv = invert_permutation(sorter)

    if method == "ordinal":
      return (inv + 1).astype(out_dtype)
    obs = jnp.concatenate([jnp.array([True]), arr[1:] != arr[:-1]])
    dense = obs.cumsum()[inv]
    if method == "dense":
      return dense.astype(out_dtype)
    count = jnp.nonzero(obs, size=arr.size + 1, fill_value=obs.size)[0].astype(out_dtype)

View on GitHub (pinned to 1e1c6a8fc0)

Solutions

Use one of: 'average', 'min', 'max', 'dense', 'ordinal' (exact lowercase).
If you need pandas-style 'first'/'min' with ordering semantics, implement via argsort or use scipy.stats.rankdata directly.

Example fix

// before
jax.scipy.stats.rankdata(x, method='first')

// after
jax.scipy.stats.rankdata(x, method='ordinal')

Defensive patterns

Strategy: validation

Validate before calling

VALID = ('average', 'min', 'max', 'dense', 'ordinal')
assert method in VALID, f"method must be one of {VALID}, got {method!r}"

Type guard

def is_valid_rank_method(m) -> bool:
    return m in ('average', 'min', 'max', 'dense', 'ordinal')

Try / catch

try:
    jax.scipy.stats.rankdata(x, method=method)
except ValueError as e:
    if 'unknown method' in str(e):
        jax.scipy.stats.rankdata(x, method='average')
    else:
        raise

Prevention

Validate config-driven method strings at load time.
Note pandas 'first' has no jax/scipy equivalent; use 'ordinal'.

When it happens

Trigger: Calling jax.scipy.stats.rankdata(a, method='dense ') with whitespace/typo, or an unsupported method name.

Common situations: Passing method from a config dict with a typo; assuming a scipy method name like 'first' (pandas rank style) exists.

Related errors

AI-assisted analysis of jax-ml/jax@1e1c6a8fc0 (2026-08-27). Data as JSON: /api/errors/00e01cbb30a3ecdc. Report an issue: GitHub.