Skip to content

Connector: data sources

Open in Colab ยท View on GitHub

Reads the same 2,000 rows from CSV, Parquet, a throwaway postgres:16-alpine container and OTLP trace exports, showing how since/limit are pushed down into SQL.

pip install pecca psycopg[binary]
jupyter lab examples/notebooks/connector_datasource.ipynb
Runtime: ~10 s. Needs Docker: yes (Postgres container; SQLite fallback otherwise). With PECCA_SKIP_DOCKER=1 (as in CI) the Docker cells are skipped and the notebook still runs end to end. Outputs are committed, so GitHub renders the results without running anything.