daft.functions.json_object_keys#

def json_object_keys(expr: Expression) -> Expression:
    """Returns the top-level keys of a JSON object as a list of strings.

    Returns ``NULL`` when the input is ``NULL``, cannot be parsed as JSON,
    or the parsed JSON is not an object. Returns an empty list for empty
    objects.

    Note:
        Keys are returned in **sorted alphabetical order**, not source
        insertion order. This differs from Spark's ``json_object_keys``,
        which preserves insertion order.

    Args:
        expr: A string expression containing JSON.

    Returns:
        Expression: A ``List[String]`` expression with the object's keys.

    Examples:
        >>> import daft
        >>> from daft.functions import json_object_keys
        >>>
        >>> df = daft.from_pydict({"col": ['{"a": 1, "b": 2}', "{}", "[1, 2]", None]})
        >>> df.with_column("keys", json_object_keys(df["col"])).collect()
        ╭──────────────────┬──────────────╮
        │ col              ┆ keys         │
        │ ---              ┆ ---          │
        │ String           ┆ List[String] │
        ╞══════════════════╪══════════════╡
        │ {"a": 1, "b": 2} ┆ [a, b]       │
        ├╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌┤
        │ {}               ┆ []           │
        ├╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌┤
        │ [1, 2]           ┆ None         │
        ├╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌┤
        │ None             ┆ None         │
        ╰──────────────────┴──────────────╯
        <BLANKLINE>
        (Showing first 4 of 4 rows)
    """
    return Expression._call_builtin_scalar_fn("json_object_keys", expr)