Join Canvas

Build a pipeline

Create a pipeline and drag tables from the connection list onto the canvas. Drag from a column to a column of another table to join (inner / left / right / full), then add Filter, Select, Group by, Union (up to 10 inputs) and Formula nodes.

The Preview panel below shows the result of the selected node; drag its edge to make it taller.

On the sample data: products (shop) joined with stock (warehouse) on sku = sku, filtered on qty_on_hand <= reorder_level, gives 24 rows.

How joins across databases work

Every side is read into a DuckDB worker. PostgreSQL, MySQL and SQLite are read in place; SQL Server, Oracle and MongoDB tables are copied (previews use the first 100,000 rows, runs at most 1,000,000 rows per table).