Help Center> Data Lake Insight> SQL Syntax Reference (To Be Offline)> Spark SQL Syntax Reference (Unavailable Soon)> Grouping> ROLLUP
Updated on 2024-01-23 GMT+08:00
View PDF

ROLLUP

Function

This statement is used to generate the aggregate row, super-aggregate row, and the total row. The statement can achieve multi-layer statistics from right to left and display the aggregation of a certain layer.

Syntax

1
2
3
 
SELECT attr_expr_list FROM table_reference
  GROUP BY col_name_list
  WITH ROLLUP;

Keyword

ROLLUP is the expansion of GROUP BY. For example, SELECT a, b, c, SUM(expression) FROM table GROUP BY a, b, c WITH ROLLUP; can be transformed into the following query statements:
  • Counting the (a, b, c) combinations 
    1
2
    		 
    SELECT a, b, c, sum(expression) FROM table
  GROUP BY a, b, c;
  • Counting the (a, b) combinations 
    1
2
    		 
    SELECT a, b, NULL, sum(expression) FROM table
  GROUP BY a, b;
  • Counting the (a) combinations 
    1
2
    		 
    SELECT a, NULL, NULL, sum(expression) FROM table
  GROUP BY a;
  • Total 
    1
    		 
    SELECT NULL, NULL, NULL, sum(expression) FROM table;

Precautions

The to-be-grouped table must exist. If this statement is used to group a table that does not exist, an error is reported.

Example

To generate the aggregate row, super-aggregate row, and total row according to the group_id and job fields and return the total salary on each aggregation condition, run the following statement:

1
2
3
 
SELECT group_id, job, SUM(salary) FROM group_test
  GROUP BY group_id, job
  WITH ROLLUP;
Parent topic: Grouping