問題文
A large fact table is usually filtered by event_date. The query profile shows that queries scan almost every micro-partition even when the filter covers a single day. Which change addresses this?
選択肢
- Increase the warehouse size so that the additional partitions can be read in parallel and the scan volume drops with it once the extra compute is active.
- Sort the table with ORDER BY and clustering in the view that the reports use, which arranges the rows for pruning.
- Define a clustering key on event_date so that each micro-partition holds a narrower range of dates and more partitions can be skipped.
- Add a secondary index on event_date so that the filter can look up the matching rows directly.