I'm loading from s3 bucket full of CSV files generated by an external pipeline (so I don't control the path structure). Is there a way to get a metacolumn into the CTAS or with another loading mechanism to include the filename/path that generates rows, akin to duckdb's read_csv_auto(..., filename=True) or to bigquery's external table that has a metacolumn _filename?

nope, only https://docs.starrocks.io/docs/sql-reference/sql-functions/table-functions/files/#columns_from_path open feature request

Just filed

Add ability to get filename/object name as returned metacolumn when loading from multiple files.

https://github.com/StarRocks/starrocks/issues/66973

NIIICE! Thanks, @Евгений Шишкин!

pushed a fix for https://github.com/StarRocks/starrocks/issues/66976 too