フリー問題

問題文

A nightly job must apply a set of changes to a target table: update the rows that match on a key and insert the rows that do not. Which approach does this in one operation with the DataFrame API?

選択肢

1. Call merge on the target table with the source DataFrame, the join expression, and a list of matched and not-matched clauses.
2. Union the source into the target and rely on the table to keep the newer row for each key.
3. Call update on the target with the source, then call an insert method for the rows that the update did not touch, in that fixed order.
4. Save the source over the target with the overwrite mode, because that keeps matching rows and adds the rest.