jax-ml/jax · error · ValueError

Need at least one array to concatenate

Error message

Need at least one array to concatenate

What it means

FragmentedArray.concatenate validates that at least one array is passed, mirroring numpy.concatenate semantics. An empty sequence has no shape, dtype, or layout to infer a result from, so the library raises ValueError immediately.

Solutions

  1. Guard the call: if not arrays: skip, return None, or construct a zero-length FragmentedArray explicitly.
  2. Fix the upstream logic so the list is guaranteed non-empty (e.g., default to a single empty array).

Example fix

# before
result = FragmentedArray.concatenate(parts, axis=0)  # parts == []
# after
result = (
    FragmentedArray.concatenate(parts, axis=0)
    if parts else None
)
Defensive patterns

Strategy: validation

Validate before calling

if not arrays:
    return  # or build an explicit zero-length fragment
result = FragmentedArray.concatenate(arrays, axis=axis)

Try / catch

try:
    out = FragmentedArray.concatenate(parts, axis=0)
except ValueError as e:
    if 'at least one array' not in str(e): raise
    out = None

Prevention

When it happens

Trigger: Calling FragmentedArray.concatenate([], axis=...) with an empty list/tuple, often because a list comprehension or accumulator produced zero arrays at runtime.

Common situations: Dynamic-length kernel pipelines where the number of arrays depends on runtime conditions (e.g., splitting work by number of tensor cores and getting 0); refactoring code that assumed a non-empty batch; passing an empty generator result.

Related errors

AI-assisted analysis of jax-ml/jax@1e1c6a8fc0 (2026-08-27). Data as JSON: /api/errors/06b04ff09426d196. Report an issue: GitHub.

Appendix: source

Thrown at jax/experimental/mosaic/gpu/fragmented_array.py:5340

  for i, dim in enumerate(src.shape):
    if dim == 1 and dst.shape[dims[i]] > 1:
      exp_indices.append(i)
    if dim > 1:
      pre_indices.append(i)
  # If both exist, all expansions must happen before all preserved
  # dimensions.
  if exp_indices and pre_indices and max(exp_indices) >= min(pre_indices):
    return False
  return True


def concatenate(
    arrays: Sequence[FragmentedArray],
    axis: int = 0,
) -> FragmentedArray:
  """Concatenates fragmented arrays along the specified axis."""
  if not arrays:
    raise ValueError("Need at least one array to concatenate")
  arr0 = arrays[0]
  rank = len(arr0.shape)
  if not -rank <= axis < rank:
    raise ValueError(f"{axis=} is out of bounds for array of {rank=}")
  if axis < 0:
    axis += rank

  if len(arrays) == 1:
    return arr0

  new_shape = list(arr0.shape)
  for i, arr in enumerate(arrays[1:], start=1):
    if len(arr.shape) != rank:
      raise ValueError(
          f"All arrays must have the same rank, got {len(arr.shape)} at index"
          f" {i} (expected {rank})"
      )
    if arr.mlir_dtype != arr0.mlir_dtype:

View on GitHub (pinned to 1e1c6a8fc0)