hard20 pts · first answer countssqlperformancewarehousing

Distinct users on billions of rows

A dashboard needs daily unique users across 20 billion events. COUNT(DISTINCT user_id) is slow and costly, and a 1 to 2 percent error is fine. What is the best approach?

A.Use COUNT(*) insteadB.Sample 10 rows per dayC.Use an approximate function such as APPROX_COUNT_DISTINCT (HyperLogLog), optionally storing HLL sketches per day to merge laterD.Export the data to a spreadsheet

← A slow Python UDF in PySpark Iceberg hidden partitioning →