Demo: database results + a local CSV

Which sales region missed its target? Revenue comes from PostgreSQL; targets come from a local CSV. Join them locally in omni-sql without uploading the CSV to the database. All values are fictional.

1. Connect to a development PostgreSQL database and run sales.sql. It uses VALUES, so it creates no tables and changes no data.
2. Choose Analyze locally and import the Full query result. A new Local DuckDB tab opens with a query on the imported table (named local_analysis in the captured release; keep whatever name your workspace assigns).
3. Choose Import file and select targets.csv. It becomes a local table named targets.
4. Run join.sql, adjusting the two table names if needed.
5. Choose Export full CSV to export the joined rows.

Expected result:

region	revenue	target	gap	target_pct
North	125000	100000	25000	125.0
West	112000	100000	12000	112.0
South	84000	100000	-16000	84.0

Regenerating the media

The README GIF and its poster are a sequence of real app screenshots kept in docs/images/release-visuals/. Regenerate them with python scripts/generate_release_visuals.py --local-analysis (needs Pillow and the Windows Segoe UI fonts). Use fictional data and a test connection, hide unrelated connections, and keep the source PNGs. If you change this demo, keep the website’s copy of the result table aligned (see website).