Function char_to_bytes

xarray/coding/strings.py:206–239

Convert numpy/dask arrays from characters to fixed width bytes.

(arr)

def char_to_bytes(arr):
    """Convert numpy/dask arrays from characters to fixed width bytes."""
    if arr.dtype != "S1":
        raise ValueError("argument must have dtype='S1'")

    if not arr.ndim:
        # no dimension to concatenate along
        return arr

    size = arr.shape[-1]

    if not size:
        # can't make an S0 dtype
        return np.zeros(arr.shape[:-1], dtype=np.bytes_)

    if is_chunked_array(arr):
        chunkmanager = get_chunked_array_type(arr)

        if len(arr.chunks[-1]) > 1:
            raise ValueError(
                "cannot stacked dask character array with "
                f"multiple chunks in the last dimension: {arr}"
            )

        dtype = np.dtype("S" + str(arr.shape[-1]))
        return chunkmanager.map_blocks(
            _numpy_char_to_bytes,
            arr,
            dtype=dtype,
            chunks=arr.chunks[:-1],
            drop_axis=[arr.ndim - 1],
        )
    else:
        return StackedBytesArray(arr)


def _numpy_char_to_bytes(arr):