Preparation node
Sample
Keep a subset of rows: the first or last N, skip N, every Nth, a 1-in-N chance, or exactly N at random.
What it does
Keeps a subset of rows without changing the schema. Choose first / last / skip N, every Nth row, a 1-in-N random chance per row, or exactly N random rows. An optional seed makes the random methods reproducible.
Inputs and outputs
| Inputs | One. |
| Outputs | One. |
When to use it
- Grab a quick preview slice
- Take a representative random sample
- Skip a header block or thin out data
Examples
First 100
First N rows → 100
Every 10th row
1 of every N rows → 10
A systematic sample.
Exactly 500 random
Random N rows → 500 (seed 42)
A seed gives the same draw each run.
1-in-20 chance
1 in N chance → 20
Result size varies.
Tips
- Set a seed on the random methods for a reproducible sample.
- Last N and Random read the whole input (they spill if large); a Test run only samples the 50-row preview.
- The schema is unchanged; only rows are dropped.