Pass S3 objects as parameters to PostgreSQL, MSSQL, MySQL, BigQuery and Snowflake scripts and read them with the dialect's JSON-table function.
New features
- Declare an `(s3object)` argument in any native SQL dialect; the worker downloads the file and binds it as a JSON parameter.
- Auto-detects format from the object key extension: `.parquet`, `.csv`, `.json`/`.jsonl`.
- Symmetric to the existing `-- s3` output flag: a previous flow step writes to S3, the next native SQL step reads it as a parameter.