pathwaycom/pathway · error · TypeError

{name} has to be a Table instead of {type(arg)}

Error message

{name} has to be a Table instead of {type(arg)}

What it means

Inside a @pw.table_transformer decorated iteration function (transformer used with pw.iterate / transform_iterate), every argument must be either a pw.Table or a value wrapped in pw.iterate_universe. The operator machinery copies each input table before invoking the iteration logic; anything else (an int, a DataFrame, None) cannot be processed as a table and raises this TypeError.

Source

Thrown at python/pathway/internals/operator.py:365

        self._universe_mapping = defaultdict(Universe)

    def __call__(self, **kwargs):
        input = as_arg_tuple(kwargs)

        input_copy = ArgTuple.empty()
        iterated_with_universe_copy = ArgTuple.empty()

        # unwrap input and materialize input copy
        for name, arg in input.items():
            if isinstance(arg, pw.Table):
                input_copy[name] = self._copy_input_table(name, arg, unique=False)
            elif isinstance(arg, iterate_universe):
                iterated_with_universe_copy[name] = self._copy_input_table(
                    name, arg.table, unique=True
                )
                input[name] = arg.table
            else:
                raise TypeError(f"{name} has to be a Table instead of {type(arg)}")

        assert all(isinstance(table, pw.Table) for table in input)

        # call iteration logic with copied input and sort result by input order
        raw_result = self.func_spec.func(**input_copy, **iterated_with_universe_copy)
        arg_tuple = as_arg_tuple(raw_result)
        result = arg_tuple.process_input(input)
        if not iterated_with_universe_copy.is_key_subset_of(result):
            raise ValueError(
                "not all arguments marked as iterated returned from iteration"
            )
        for name, table in result.items():
            input_table: pw.Table = input[name]
            assert isinstance(table, pw.Table)
            input_schema = input_table.schema._dtypes()
            result_schema = table.schema._dtypes()
            if input_schema != result_schema:
                raise ValueError(

View on GitHub (pinned to fa2f74a464)

Solutions

1. Pass only pw.Table objects (or pw.iterate_universe(pw.Table) wrappers) as transformer arguments; capture constants in a closure or functools.partial instead.
2. Convert pandas DataFrames with pw.debug.table_from_pandas() or io via pw.io.csv.read before calling the transformer.
3. If you meant a table carrying its universe through iteration, wrap it: pw.iterate_universe(table).

Example fix

# before
@pw.table_transformer
def scale(t: pw.Table, factor):  # factor: int reaches the operator
    ...
scale(table, 2)

# after
@pw.table_transformer
def scale(t: pw.Table, factor: float = 2.0):
    return t.select(v=pw.this.v * factor)
scale(table)

import pathway as pw

def all_args_are_tables(kwargs: dict) -> bool:
    return all(isinstance(v, (pw.Table, pw.iterate_universe)) for v in kwargs.values())

import pathway as pw
from typing import TypeGuard

def is_table(v) -> TypeGuard[pw.Table]:
    return isinstance(v, pw.Table)

When it happens

Trigger: Defining a @pw.table_transformer function whose parameter is not typed as pw.Table and calling it (directly or via pw.iterate) with a non-Table value, e.g. def my_transformer(t: pw.Table, factor: int) where factor is passed positionally into the iteration machinery, or passing a pandas DataFrame instead of a pw.Table.

Common situations: Mixing pandas/polars DataFrames with Pathway tables, passing config scalars as transformer arguments instead of closing over them, or passing pw.this.column (a ColumnReference) where a whole table is required.

Related errors

• not all arguments marked as iterated returned from iteration
• output: {result_schema} of the iterated function does not c
• iterated table not marked as 'iterate_universe' changed its
• direction argument of join should be of type asof_join.Direc
• The interval argument of a join should be of a type pathway.

AI-assisted analysis of pathwaycom/pathway@fa2f74a464 (2026-08-15). Data as JSON: /api/errors/8aa9d4d1e6fcb221. Report an issue: GitHub.