Output Writing

Output Writing

This page documents how output paths are constructed and how files are written.

Output path selection

The writer chooses the output path based on these inputs:

• --dir (optional output directory)
• --output (optional output file name or list)
• --format (file type/extension)
• ticker, period, timeframe (for auto-naming)

Auto-generated naming

If --output is not provided, the output file name is:

<ticker>_<period>_<timeframe>.<ext>

Examples:

• AAPL_1y_1d.csv
• MSFT_2y_1wk.parquet

With --dir

If --dir is provided, the file is written inside that directory. The directory is created if it does not exist.

Format writers

The tool writes based on the final extension:

• .csv → polars.DataFrame.write_csv
• .parquet → write_parquet
• .json → write_json
• .xlsx → write_excel
• .avro → write_avro

Concurrency model

Writes are scheduled as asyncio tasks, but the actual write calls run in threads via asyncio.to_thread(...).

This helps avoid blocking the event loop while writing large files.

Related pages:

• Write Output Deep Dive
• Output Formats
• Advanced Usage
• Pipeline