Output
PostgreSQL Output
Writes the rows that reach it into an existing PostgreSQL table, in one transaction, and passes them on unchanged.
When to use it
Use it to land cleaned or staged rows in PostgreSQL — a reporting table, a staging table for a downstream job.
In a pipeline
File InputType ConversionPostgreSQL Output
Settings
| Setting | Type | Required | Default or hint |
|---|---|---|---|
| Connection | select | Yes | Add a connection below |
| Target Table | select | Yes | Select a table |
| Write Mode | select | — | append |
| Key columns | select | — | Columns that identify a row |
Example
Land staged rows
- Target Table=staging.orders
- Write Mode=Append rows
| order_id | total |
|---|---|
| 1042 | 860 |
| 1043 | 312.5 |
| order_id | total |
|---|---|
| 1042 | 860 |
| 1043 | 312.5 |
Worth knowing
- Rows go into the columns whose names match; other columns are ignored, and __ columns are never written.
- Append rows inserts; Replace existing rows empties the table first (DELETE, inside the same transaction) — if any insert fails, nothing changes.
- Previews never write: they say how many rows would be written.
- Upsert updates the rows whose Key columns already exist and adds the rest — native INSERT … ON CONFLICT when the keys have a unique index or primary key, otherwise the incoming keys are deleted first, in the same transaction.