GROUP BY Clause - Teradata Database

GROUP BY Clause

Purpose  

Groups result rows by the values in one or more columns or by various extended GROUP BY operations on specified column expressions.

Syntax  

where:

ANSI Compliance

The GROUP BY clause is ANSI SQL:2011-compliant with extensions.

GROUP BY Clause Terminology

A GROUP BY clause is said to be simple if it does not contain any of the following elements:

A GROUP BY clause is said to be primitive if it does not contain any of the following elements:

See “GROUPING SETS Option” on page 157 for definitions of rollup_list, cube_list, grouping_sets_specification, and empty_grouping_set.

See the following topics for examples of non‑simple, non‑primitive GROUP BY clauses:

Ordinary Grouping Set Expressions

The following table provides the definitions for the valid ordinary grouping set expressions you can specify in a GROUP BY clause.

Using the GROUP BY Clause

The rules and restrictions are:

For window functions, such as SUM and AVG, GROUP BY collapses all rows with the same value for the group-by columns into a single row. The GROUP BY clause must include all the columns specified in the:

For examples and more information on GROUP BY and ordered analytical functions, see SQL Functions, Operators, Expressions, and Predicates.

a WHERE

b GROUP BY

c HAVING

Reason for Unexpected Row Length Errors: Sorting Rows for Grouping

Before performing the sort operation that groups the rows to be returned to the requestor, Teradata Database creates a sort key and appends it to the rows to be sorted. If the length of this temporary data structure exceeds the system row length limit of 64 KB, the operation returns an error to the requestor. Depending on the situation, the message text is one of the following:

For explanations of these messages, see Messages.

The BYNET only looks at the first 4096 bytes of the sort key created to sort the specified fields, so if the field the sort key is based on is greater than 4096 bytes, the key is truncated and the data might or might not come back in the desired order.

Examples

These examples show the use of the simple or primitive GROUP BY clause.

Examples for the ROLLUP, CUBE, and GROUPING SETS extended GROUP BY operators are provided in “ROLLUP Grouping Set Option” on page 160, “CUBE Grouping Set Option” on page 154, and “GROUPING SETS Option” on page 157, respectively.

Example : Simple GROUP BY Operation

Generate a report of salary totals by department, the result might look something like the report that follows.

     SELECT dept_no, SUM(salary) 

     FROM employee 

     GROUP BY dept_no;

The result might be similar to the report that follows.

Example : Specifying a GROUP BY Clause on Nonaggregate Expressions When the Select List Includes an Aggregate

If you specify an aggregate in the select expression list of a query, then you must also specify a GROUP BY clause that includes all nonaggregate expressions from the select list. Otherwise, the system returns the following message.

     Selected non‑aggregate values must be part of the associated group.

The system returns error message 3504 whenever an aggregate query includes a nonaggregate expression in its SELECT list, WHERE clause, ORDER BY clause, or HAVING clause, but not in a GROUP BY clause.

The system also returns this error when ORDER BY and WITH clauses specify aggregates, but the select list for the query does not.

For example, Teradata Database aborts the following query because it does not specify a GROUP BY clause that groups on the only nonaggregate expression in the select list, which is department_number:

     SELECT department_number, SUM(salary_amount)

     FROM employee

     WHERE department_number IN (100, 200, 300);

To work as intended, the query must be rewritten with an appropriate GROUP BY clause:

     SELECT department_number, SUM(salary_amount)

     FROM employee

     WHERE department_number IN (100, 200, 300)

     GROUP BY department_number;

The following statement aborts and returns an error because it does not specify all of the nonaggregate columns from the select list in its GROUP BY clause.

     SELECT employee.dept_no, department.dept_name, AVG(salary)

     FROM employee, department

     WHERE employee.dept_no = department.dept_no

     GROUP BY employee.dept_no;

In this case, the qualified nonaggregate column department.dept_name is missing from the GROUP BY clause.

The following statement aborts and returns an error because the nonaggregate grouping column specified in the ORDER BY clause is not also specified in the GROUP BY clause.

     SELECT employee.dept_no, AVG(salary)

     FROM employee, department

     WHERE employee.dept_no = department.dept_no

     ORDER BY department.dept_name

     GROUP BY employee.dept_no;

The following statement, based on the table definitions shown, aborts and returns an error because the aggregate query includes a nonaggregate expression, d1, in its HAVING clause, but not in the GROUP BY clause.

     CREATE TABLE t1(a1 int, b1 int, c1 int, d1 int);

     CREATE TABLE t2(a2 int, b2 int, c2 int, d2 int);

     SELECT min(a1) as i, max(b1) as j from t1

     GROUP BY c1

     HAVING 30 >= (sel count(*) from t2 where t1.d1=5); 

The correct form of the query includes the nonaggregate expression, d1, in its HAVING clause and in the GROUP BY clause.

     SELECT min(a1) as i, max(b1) as j from t1

     GROUP BY c1, d1

     HAVING 30 >= (sel count(*) from t2 where t1.d1=5);

Example : Specifying GROUP BY Using an Ordered Analytic Function

The following statement specifies a GROUP BY clause with an ordered analytical function to generate report breaks where the function resets and computes a new value for the next grouping.

The example groups all items into percentile by profitability for each store and then returns only the items of interest, which, in this case, are the lowest percentile for each store.

     SELECT store, item, profit, QUANTILE(100, profit) AS percentile

     FROM (SELECT items.item, SUM(sales) -

                  (COUNT(sales)*items.item_cost) AS profit

           FROM daily_sales, items

           WHERE daily_sales.item = items.item

           GROUP BY items.item,items.itemcost) AS item_profit

     GROUP BY store, item, profit, percentile

     QUALIFY percentile = 99;

The result of this query looks like the following table:

store

-----

item

----

profit

------

percentile

----------

Eastside

Golf balls

    100.19

99

Westside

Tennis balls

   -110.00

99

Central

Bowling balls

   -986.81

99

South

Codfish balls

- 1,891.89

99

North

Bocce balls

  1,180.88

99

Example : SELECT Statement With a Scalar Subquery in Its GROUP BY Clause

The following example specifies a scalar subquery in its GROUP BY clause:

     SELECT sale_date, SUM(amount) 

     FROM sales_table AS s

     GROUP BY sale_date, (SELECT prod_name 

                          FROM prod_table AS p 

                          WHERE p.prod_no = s.prod_no);

Example : GROUP BY and PERIOD Value Expressions

The following example shows how the GROUP BY clause operates on PERIOD value expressions, where period_of_stay is the PERIOD value expression.

     SELECT emp_no, period_of_stay 

     FROM employee 

     GROUP BY emp_no, period_of_stay;

For More Information

For more information related to the GROUP BY clause, see: