問題文
A 3 TB fact table is almost always filtered on a high-cardinality order identifier column. Queries are slow because they read far more data than they need. Which data layout technique is recommended for this table?
選択肢
- Increasing the warehouse size, because reading less data is a function of how much compute is available.
- Creating a standard view that applies the usual filter, so consumers read less data by default.
- Liquid clustering on the filtered column, so that files are organized by that column and unrelated files can be skipped.
- Partitioning by the order identifier, so that each distinct identifier gets its own directory and can be addressed directly.