ANY_VALUE¶

Syntax¶

Aggregate function

ANY_VALUE( [ DISTINCT ] <expr1> )

Window function

ANY_VALUE( [ DISTINCT ] <expr1> ) OVER ( [ PARTITION BY <expr2> ] )

Usage notes¶

• The DISTINCT keyword can be specified for this function, but it does not have any effect.

• When this function is called as a window function, it does not support:

  • An ORDER BY clause within the OVER clause.

  • Explicit window frames.

Using ANY_VALUE with GROUP BY statements¶

ANY_VALUE can simplify and optimize the performance of GROUP BY statements. A common problem for many queries is that the result of a query with a GROUP BY clause can only contain expressions used in the GROUP BY clause itself, or results of aggregate functions. For example:

SELECT customer.id , customer.name , SUM(orders.value)
    FROM customer
    JOIN orders ON customer.id = orders.customer_id
    GROUP BY customer.id , customer.name;

In this query, the customer.name attribute needs to be in the GROUP BY to be included in the result. This is unnecessary (e.g. when customer.id is known to be unique) and makes the computation possibly more complex and slower. Another option is to use an aggregate function. For example:

SELECT customer.id , MIN(customer.name) , SUM(orders.value)
    FROM customer
    JOIN orders ON customer.id = orders.customer_id
    GROUP BY customer.id;

This simplifies the GROUP BY clause, but still requires computing the MIN function, which incurs an extra cost.

With ANY_VALUE, you can execute the following query:

SELECT customer.id , ANY_VALUE(customer.name) , SUM(orders.value)
    FROM customer
    JOIN orders ON customer.id = orders.customer_id
    GROUP BY customer.id;