Concat Rows (Union / Concat) — concatRows
Stack multiple inputs row-wise (no config). This is the variadic node — connect two or more upstream nodes to its input handle and they're concatenated in order.
Use cases
- Combine monthly files into one dataset.
- Append a new batch to an existing one before further cleaning.
What it does
Stacks all connected inputs row-wise in connection order. Column alignment happens by name: inputs with different schemas produce null for missing columns.
| region | amount |
|---|---|
| North | 100 |
| South | 80 |
| region | amount |
|---|---|
| North | 100 |
| South | 80 |
| East | 150 |
| West | 90 |
TIP
The Before pane above shows Input A; the After pane shows the stacked result
after appending Input B (East/150, West/90). Connect as many inputs as you need.
Configuration
This node has no configuration — it stacks whatever inputs are connected.
Generated Python code
These exports stack df_1 then df_2 with no configuration. Columns align by name;
missing fields become null and numeric types can be relaxed. Both duplicate rows and
source connection order are retained. The upstream inputs and the engine alias must
already be defined.
pandas
df_3 = pd.concat([df_1, df_2], ignore_index=True)Polars
df_3 = pl.concat([df_1, df_2], how='diagonal_relaxed')Tips & common mistakes
- Align columns first. Inputs with different column names produce sparse rows (missing columns become null) — use Rename columns / Select columns to match schemas.
- Concat stacks rows; to combine columns on a key use Join.
- Use Remove duplicates afterward if sources overlap.