DuckDB: Speed Up String GROUP BYs With Narrow Dimension Keys
DuckDB explains how to replace repeated string group keys with sorted, narrow integer keys in a dimension table, then join labels back after aggregation. The walkthrough uses a 380,959-row dataset with 537 station names and details string hashing, comparisons, hash-table width, and perfect-hash selection.
🔗 duckdb.org
#DuckDB #SQL #QueryPerformance
Post #325
63
