GitHub
08/04/2026, 12:55 PMWe know quite a bit about the data being loaded if it was managed by Singer. We could have utilities to generate staging tables as accelerators or templates (meaning user intervention is needed).This could just build our files for us with a best guess baseline which at least accelerates development a bit. Things like casting, deduping (we know primary keys usually), etc. Another main goal of staging is renaming columns, which is a little more subjective. We could have some sort of interactive CLI like when you add a custom extractor where default is to keep the current name or the user could rename it, then that compiles to a sql model file. meltano/meltano
GitHub
08/04/2026, 12:55 PM