medium10 pts · first answer countsdata-pipelinesdeduplicationsql

Keeping one row per order

Staging has duplicate rows per order_id, sometimes with the same updated_at. Which query keeps exactly one row per order, the latest, the same way on every run?

SELECT * FROM (
  SELECT *, ROW_NUMBER() OVER (
    PARTITION BY order_id
    ORDER BY ???
  ) AS rn
  FROM staging
) WHERE rn = 1

A.ORDER BY updated_at ASC, ingest_seq ASCB.ORDER BY order_id DESC, updated_at ASCC.ORDER BY RANDOM(), updated_at DESCD.ORDER BY updated_at DESC, ingest_seq DESC

Discussion 0

Join free to post in the discussion.

Keeping one row per order

Discussion 0

Related questions