multi_store_writer

sqllocks_spindle.output.multi_store_writer

MultiStoreWriter — write Spindle output to multiple targets concurrently.

Wraps any combination of Spindle writers (DeltaWriter, WarehouseBulkWriter, FabricSqlDatabaseWriter, EventhouseWriter, etc.) and fans out write calls to all of them in parallel via ThreadPoolExecutor.

Example::

from sqllocks_spindle.output import DeltaWriter, MultiStoreWriter
from sqllocks_spindle.fabric import WarehouseBulkWriter

writer = MultiStoreWriter([
    DeltaWriter(output_dir="./delta"),
    WarehouseBulkWriter(connection_string="..."),
])
results = writer.write_all(result.tables)

Classes

SpindleWriter

Bases: Protocol

Protocol for any Spindle output writer that supports write_all().

MultiStoreResult dataclass

Aggregated result from a MultiStoreWriter.write_all() call.

MultiStoreWriter

Fan out Spindle table writes to multiple writers in parallel.

Parameters:

Name	Type	Description	Default
writers	list[SpindleWriter]	List of writer instances. Each must implement write_all().	required
max_workers	int | None	Max parallel writer threads. Defaults to number of writers.	None
raise_on_error	bool	If True, re-raise the first writer exception after all writes complete. If False (default), collect errors in the result.	False

Methods:

write_all(tables, **kwargs)

Write all tables to every configured writer in parallel.

Parameters:

Name	Type	Description	Default
tables	dict[str, DataFrame]	Table name -> DataFrame mapping (e.g. result.tables).	required
**kwargs	Any	Forwarded to each writer's write_all() call.	{}

Returns:

Type	Description
MultiStoreResult	MultiStoreResult with per-writer results and any errors.