wiring.get_error_rows_path()

Return the path of the parquet file build_file_writer writes an asset’s error rows to.

Usage

Source

wiring.get_error_rows_path(
    error_rows_key,
    error_rows_dir,
    partition_key=None,
)

The layout matches UPathIOManager’s, so with error_rows_dir set to a manager’s base_dir, the file is beside the asset’s own file.

Parameters

error_rows_key: dg.AssetKey

The key the writer writes under, such as the one get_error_rows_key returns.

error_rows_dir: UPath | Path | str

Any path UPath accepts, such as a local directory or s3://bucket/prefix. Credentials for a remote path come from the environment.

partition_key: str | None = None

The partition the rows belong to, or None for an unpartitioned asset. A dg.MultiPartitionKey becomes its dimension keys joined by /, in dimension-name order. Escaping a .. and a leading / keeps the file under error_rows_dir.

Returns

UPath

<error_rows_dir>/<key part>/.../<last part>.parquet, or <error_rows_dir>/<key part>/.../<last part>/<partition>.parquet for a partition. get_error_rows_path creates no parent directories.

Examples

import dagster as dg

import dagster_dataframely as dd

dd.wiring.get_error_rows_path(
    dd.wiring.get_error_rows_key(dg.AssetKey(["sales", "orders"])),
    "/data",
    "2026-08-01",
)

PosixUPath('/data/sales/orders__error_rows/2026-08-01.parquet')