sql “交叉应用”版本的mysql

Posted 2021-05-13

 8
down vote
accepted
	

Your closest direct approximation is a join with a correlated sub-query as the predicate.

SELECT
   ORD.ID
  ,ORD.NAME
  ,ORD.DATE
  ,ORD_HISTORY.VALUE
FROM
  ORD
INNER JOIN
  ORD_HISTORY
    ON  ORD_HISTORY.<PRIMARY_KEY>
        =
        (SELECT ORD_HISTORY.<PRIMARY_KEY>
           FROM ORD_HISTORY
          WHERE ORD.ID = ORD_HISTORY.ID
            AND ORD.DATE <= ORD_HISTORY.DATE
       ORDER BY ORD_HISTORY.DATE DESC
          LIMIT 1
        )

In your case, however, you only need one field from the target table. This means that you are able to use the correlated sub-query directly in the SELECT statement.

SELECT
   ORD.ID
  ,ORD.NAME
  ,ORD.DATE
  ,(SELECT ORD_HISTORY.VALUE
      FROM ORD_HISTORY
     WHERE ORD.ID = ORD_HISTORY.ID
       AND ORD.DATE <= ORD_HISTORY.DATE
  ORDER BY ORD_HISTORY.DATE DESC
     LIMIT 1
   )   AS VALUE
FROM
  ORD