来自Select的Oracle Update - 同一个表

Question

我正在尝试使用date2，date3，date4列中的最高日期更新列date1。我有选择声明

SELECT GREATEST(
    COALESCE(date2, date3, date4),
    COALESCE(date3, date4, date2),
    COALESCE(date4, date2, date3)
)
FROM (
    SELECT date1,date2,date3,date4 
    FROM my_table
    WHERE date1 is not null and (date1 < date2 or date1 < date3 or date1 < date4);

但我不知道如何使用它与更新或合并？

+------------------+------------------+------------------+------------------+
|      date1       |      date2       |      date3       |      date4       |
+------------------+------------------+------------------+------------------+
| 2017-04-13 16:54 | 2017-04-13 16:57 |                  | 2016-06-16       |
| 2017-04-13 15:41 |                  | 2017-04-13 15:42 | 2016-06-16       |
| 2017-04-13 15:43 | 2018-01-10 14:23 | 2017-04-13 15:41 |                  |
| 2017-04-13 16:05 |                  | 2017-04-13 16:05 | 2016-06-16       |
| 2017-04-13 16:43 | 2017-04-13 16:43 |                  | 2016-06-16       |
| 2017-04-13 16:52 | 2017-04-13 16:52 | 2017-04-13 16:07 | 2017-04-17 16:07 |
| 2018-01-10 14:20 |                  | 2018-01-10 14:23 |                  |
| 2017-09-27 14:54 | 2015-09-08 09:56 | 2017-09-27 14:54 | 2015-03-13       |
| 2017-06-16 13:38 |                  | 2017-06-16 13:39 |                  |
+------------------+------------------+------------------+------------------+

更新/合并后我想要这样的东西

+------------------+------------------+------------------+------------------+
|      date1       |      date2       |      date3       |      date4       |
+------------------+------------------+------------------+------------------+
| 2017-04-13 16:57 | 2017-04-13 16:57 |                  | 2016-06-16       |
| 2017-04-13 15:42 |                  | 2017-04-13 15:42 | 2016-06-16       |
| 2018-01-10 14:23 | 2018-01-10 14:23 | 2017-04-13 15:41 |                  |
| 2017-04-13 16:05 |                  | 2017-04-13 16:05 | 2016-06-16       |
| 2017-04-13 16:43 | 2017-04-13 16:43 |                  | 2016-06-16       |
| 2017-04-17 16:07 | 2017-04-13 16:52 | 2017-04-13 16:07 | 2017-04-17 16:07 |
| 2018-01-10 14:23 |                  | 2018-01-10 14:23 |                  |
| 2017-09-27 14:54 | 2015-09-08 09:56 | 2017-09-27 14:54 | 2015-03-13       |
| 2017-06-16 13:39 |                  | 2017-06-16 13:39 |                  |
+------------------+------------------+------------------+------------------+

Answer 1

你应该能够做到：

update t
    set date1 = greatest(date1, coalesce(date2, date1), coalesce(date3, date1), coalesce(date4, date1))
    where date1 < greatest(date1, coalesce(date2, date1), coalesce(date3, date1), coalesce(date4, date1));

在您的数据中，date1似乎永远不会是NULL。