Suppose you have the following data:
+-------+--------+------------+ | Name | Rating | Date | +-------+--------+------------+ | Alice | 4.5 | 01/01/2022 | | Alice | 4 | 14/12/2021 | | Alice | 4 | 16/05/2021 | | Mary | 5 | 05/01/2022 | | Mary | 4 | 31/01/2022 | | Bob | 3.5 | 03/02/2022 | +-------+--------+------------+
What would be the best way to group by name
and return the row with the most recent date
in Snowflake (ANSI SQL)? Expected output:
+-------+--------+------------+ | Name | Rating | Date | +-------+--------+------------+ | Alice | 4.5 | 01/01/2022 | | Mary | 4 | 31/01/2022 | | Bob | 3.5 | 03/02/2022 | +-------+--------+------------+
Advertisement
Answer
With QUALIFY you can keep the newest per name
SELECT * FROM table QUALIFY row_number() OVER (PARTITION BY name ORDER BY date desc) = 1;
As you will see in the doc’s it’s the same as Tim’s answer without the need for the nested select.