How to Find the First Row in Each Group Using PostgreSQL's DISTINCT ON Clause?

Find the first row in each grouping using PostgreSQL’s DISTINCT ON clause

Introduction

When processing grouped data, you usually need to extract the first row in each grouping. This task can be accomplished using the DISTINCT ON clause, a PostgreSQL-specific extension.

Query structure

SELECT DISTINCT ON (grouping_expression)
       selected_columns
FROM   table_name
ORDER  BY grouping_expression, additional_ordering_expression(s)

grouping_expression: Expression used to define grouping. The leading expression in this clause must match the expression in the following ORDER BY clause.

selected_columns: The columns to retrieve in the output.

additional_ordering_expression(s): Optional expression used to determine the order of rows within each grouping. These expressions must follow the grouping expression in the ORDER BY clause.

Example

Consider the purchases table with the following schema:

CREATE TABLE purchases (
    id SERIAL PRIMARY KEY,
    customer VARCHAR(50) NOT NULL,
    total NUMERIC(10, 2) NOT NULL
);

and the following data:

INSERT INTO purchases (customer, total) VALUES
('Joe', 5),
('Sally', 3),
('Joe', 2),
('Sally', 1);

To retrieve the customer with the highest total purchase amount, use the following query:

SELECT DISTINCT ON (customer)
       id, customer, total
FROM   purchases
ORDER  BY customer, total DESC, id;

Output:

<code>id | customer | total
-----------------------------
1  | Joe      | 5
2  | Sally    | 3</code>

NULL value

If the grouped value may contain NULL values, it is recommended to add ORDER BY to the NULLS LAST clause to ensure that rows with non-NULL values ​​are processed first:

ORDER  BY customer, total DESC NULLS LAST, id;

Select another column

SELECTLists are not limited to grouping expressions. Any additional columns can be included, making this technique useful for replacing complex subqueries or aggregate/window functions.

Performance Notes

DISTINCT ONThe efficiency of the query depends on the distribution of data and the number of rows in each grouping. It is usually very efficient for small groupings. However, for large groupings, the alternative technique described in the following link may provide better performance:

• SELECT DISTINCT in PostgreSQL is slower than expected
• Optimize the GROUP BY query to retrieve the latest row for each user
• Query the last N related rows of each row

The above is the detailed content of How to Find the First Row in Each Group Using PostgreSQL's DISTINCT ON Clause?. For more information, please follow other related articles on the PHP Chinese website!