wrds-psql

Query WRDS data locally using psql with .pgpass service file authentication, then pipe output through a Python pipeline to generate Parquet files and metadata.json. This enforces single-line psql commands and returns structured results suitable for analysis across datasets like CRSP and Compustat.

The best way to extract large WRDS datasets via PostgreSQL is to chunk queries by date ranges, use the COPY command for bulk export, and validate data with minimal intermediate files. This approach maintains efficiency when pulling from OptionMetrics or Compustat.

Yes, local psql connections to WRDS require a .pgpass service file for authentication. You also need a standard PostgreSQL client installed locally to execute the deterministic query pipeline and export data.

Yes, CSV export is supported alongside Parquet. The pipeline defaults to Parquet with metadata, piping psql output to Python to generate both Parquet files and metadata.json, but you can also extract CSV outputs depending on your analysis needs.

WRDS psql commands must be on a single line because the skill enforces single-line psql commands to ensure deterministic execution within the local PostgreSQL service file pipeline. This prevents parsing errors during automated data extraction.

This approach supports both inline SQL and file-based queries for WRDS data extraction. The deterministic pipeline processes either query pattern using local psql connections and exports structured results to Parquet with metadata.