Keep only the rows you need

Change one value, run the pipeline, and watch how a filter transforms the orders dataset.

transformations.yamlchange one value

transformations:
  steps:
    - filter:
        expr: amount > 20
        # keep matching rows
        on_error: stop

Minimum order amount€20

Your goal

Keep orders above €50. Move the threshold, then run the pipeline.

orders.csv→filter→preview

Output preview8 input rows

order_id	amount	status

Simulation ready · no data leaves your browser

That’s it — 2 rows remain.You changed the filter condition and verified its effect on real sample rows.

filter

Generated from the engine models.

1 key

key	type	required	default	constraint
expr	string	yes	—	au moins 1 caractere(s)

What just happened?

Hydra evaluates the expression on every row. A row whose amount is empty never matches — a comparison against a missing value is false, not an error.

Good to know

On a compatible SQL source, Hydra can push this filter down as a WHERE clause instead of loading every row.