Hot partition skew

Return partition_day, row_count, and skew_flag ('hot' when row_count is above the average, else 'normal'). Order by row_count DESC, then partition_day.

Label each partition hot or normal by comparing its row_count to the table average.

Sample input

table_partitionsAverage row count is 1625; two days sit well above it.

partition_day	row_count
2026-01-01	900
2026-01-02	1100
2026-01-03	800
2026-01-04	4200
2026-01-05	1000
2026-01-06	3000
2026-01-07	1200
2026-01-08	800

Expected output

Expected outputEach partition flagged hot or normal versus the average, largest first.

partition_day	row_count	skew_flag
2026-01-04	4200	hot
2026-01-06	3000	hot
2026-01-07	1200	normal
2026-01-02	1100	normal
2026-01-05	1000	normal
2026-01-01	900	normal
2026-01-03	800	normal
2026-01-08	800	normal

Constraints

Compare row_count to (SELECT AVG(row_count) FROM table_partitions) inside a CASE to label hot vs normal. Order by row_count DESC, partition_day ASC.

Expected skills

Skew detection with a scalar AVG and CASE labels.

SQL

AI evaluation

Run the SQL query to inspect preview rows.