Overview
Setting up a database replication previously meant typing schema and table names by hand, writing the extraction SQL yourself, and finding out whether any of it was right on the next workflow run. Connection failures returned a generic message with nothing to act on. This release rebuilds that flow for all database sources - Azure, BigQuery, SQL Server, MongoDB, MySQL, NetSuite, PostgreSQL, Redshift, Salesforce Service Cloud, and Snowflake.
What we changed
Source Schema and Source Table are now dropdowns read live from your database, and Fetch Columns lists every column so you can choose which to replicate, rename destinations, and set data types. Because columns are selected explicitly, the generated query names each one rather than using SELECT * - adding a column to your source table no longer breaks the extraction. Four filtering tabs replace hand-written SQL, with a Generated SQL Query panel showing exactly what will run. Test Query runs that query against your source, limited to 5 rows, with results downloadable as CSV, and Test Connection now returns the actual failure rather than a generic one. Editing is scoped to Name, Action, and Sync Key, with a prompt protecting your destination table when a sync key changes; deleting now asks whether to keep, drop, or rename it.
What to expect
The update rolls out automatically and no action is required. Existing Table Replicators keep running unchanged - you'll find them on the Integrations page. Database Replicators covers the full flow, including a video walkthrough of creating one.
Equivalent CSV and file import improvements are anticipated in early September.