如何覆盖 SQL 中的 row_number 值？

Question

如何覆盖 SQL (Redshift) 中的列？

以下查询给了我一个错误Target table must be part of an equijoin predicate

UPDATE table1
SET rank = temp.new_rank
FROM (
      SELECT ROW_NUMBER() OVER (PARTITION BY session_id ORDER BY date) AS new_rank
      FROM table1
      ) temp;

Answer 1

您需要一个WHERE子句将表的每一行与子查询的一行匹配：

UPDATE table1 AS t1
SET rank = t.new_rank
FROM (
  SELECT *, ROW_NUMBER() OVER (PARTITION BY session_id ORDER BY date) AS new_rank
  FROM table1
) AS t
WHERE t.session_id = t1.session_id AND t.date = t1.date;

我在WHERE子句中使用列session_id和date ，但如果有另一列，如唯一 id 或任何其他主键，最好使用它。