pathwaycom/pathway · error · ValueError

'topic_name' must be a non-empty string; got an empty string

Error message

'topic_name' must be a non-empty string; got an empty string. Kafka does not allow empty topic names.

What it means

pw.io.kafka.write rejects an empty-string topic_name because Kafka does not allow zero-length topic names. Note this check only covers str inputs (isinstance(topic_name, str) and not topic_name) — a non-string topic_name is passed through and handled downstream, e.g. a ColumnReference is allowed for per-row topic routing.

Source

Thrown at python/pathway/io/kafka/__init__.py:717

    >>> pw.io.kafka.write(
    ...     t2,
    ...     rdkafka_settings,
    ...     "test",
    ...     format="raw",
    ...     key=t2.bar,
    ...     value=t2.foo,
    ...     headers=[t2.baz],
    ... )
    """
    if not rdkafka_settings.get("bootstrap.servers"):
        raise ValueError(
            "rdkafka_settings must contain a non-empty 'bootstrap.servers' "
            "entry so the producer can locate a broker; got "
            f"{rdkafka_settings.get('bootstrap.servers')!r}."
        )
    if isinstance(topic_name, str) and not topic_name:
        raise ValueError(
            "'topic_name' must be a non-empty string; got an empty string. "
            "Kafka does not allow empty topic names."
        )

    output_format = MessageQueueOutputFormat.construct(
        table,
        format=format,
        delimiter=delimiter,
        key=key,
        value=value,
        headers=headers,
        topic_name=topic_name if isinstance(topic_name, ColumnReference) else None,
        schema_registry_settings=schema_registry_settings,
        subject=subject,
    )
    output_table = output_format.table
    remapped_sort_by = _remap_sort_by(sort_by, table, output_table)

View on GitHub (pinned to fa2f74a464)

Solutions

1. Pass a concrete topic name: pw.io.kafka.write(table, rdkafka_settings, 'test', format='raw').
2. Default config values at load time: topic_name = os.environ.get('KAFKA_TOPIC') or 'test'.
3. Kafka topic names must match [a-zA-Z0-9._-]{1,255} — validate user-supplied names against that pattern.

Example fix

# before
topic_name = os.environ.get("KAFKA_OUTPUT_TOPIC")  # unset -> None/''
pw.io.kafka.write(t, rdkafka_settings, topic_name or "", format="raw")
# after
topic_name = os.environ.get("KAFKA_OUTPUT_TOPIC") or "test"
pw.io.kafka.write(t, rdkafka_settings, topic_name, format="raw")

if isinstance(topic_name, str) and not topic_name:
    raise SystemExit("topic_name must be a non-empty Kafka topic name")

pw.io.kafka.write(table, rdkafka_settings, topic_name, format="raw")

def valid_write_topic(v) -> bool:
    import re
    if isinstance(v, str):
        return bool(re.match(r"^[a-zA-Z0-9._-]{1,255}$", v))
    return v is not None  # ColumnReference allowed for per-row routing

When it happens

Trigger: pw.io.kafka.write(table, rdkafka_settings, '', format='raw') — literal empty string; or topic_name computed from config that resolved to ''.

Common situations: Topic name from an unset env variable defaulting to ''; whitespace-only or accidentally cleared config value; templated deployment where the topic variable was not interpolated. Remember write uses 'topic_name' while read uses 'topic'.

Related errors

• Failed to install dependencies
• Column {api.TIME_PSEUDOCOLUMN} cannot contain negative times
• 'subject' must be a non-empty string; got an empty string. S
• 'topic' must be a non-empty string; got {topic!r}. Kafka doe
• Column {pseudocolumn} has to contain integers only.

AI-assisted analysis of pathwaycom/pathway@fa2f74a464 (2026-08-15). Data as JSON: /api/errors/46b26f837e75d89c. Report an issue: GitHub.