Han*_*Gay 5 oracle ddl timezone timestamp data-migration
我有一对列不幸被错误地定义为TIMESTAMP(6)而不是TIMESTAMP(6) WITH TIME ZONE.我想将这些列从旧的错误数据类型迁移到新的,正确的数据类型.最重要的是,值似乎已在E(S | D)T中捕获,我需要UTC中的值.
到目前为止,我得到的最好的是:
alter table OOPSIE_TABLE add (
NEW_COLUMN_A timestamp(6) with time zone,
NEW_COLUMN_B timestamp(6) with time zone
);
update OOPSIE_TABLE set
NEW_COLUMN_A = COLUMN_A,
NEW_COLUMN_B = COLUMN_B
;
alter table OOPSIE_TABLE drop column (
COLUMN_A,
COLUMN_B
);
alter table OOPSIE_TABLE rename column NEW_COLUMN_A to COLUMN_A;
alter table OOPSIE_TABLE rename column NEW_COLUMN_B to COLUMN_B;
Run Code Online (Sandbox Code Playgroud)
不幸的是,这给我留下了15-JUN-12 05.46.29.600102000 PM -04:00我想要的数据15-JUN-12 09.46.29.600102000 PM UTC(或者Oracle会将其格式化).
我已经完成select dbtimezone from dual;并且它显示了我+00:00,所以我不确定如何继续.理想情况下,我可以在纯DML中执行此操作,并根据旧日期值(我确定在America/New_York时区中)将其计入DST.
在@JustinCave 的帮助下,我得到了以下解决方案,这正是我想要的:
-- Rename the old columns so we can use them as a data source *AND* so
-- we can roll back to them if necessary.
alter table OOPSIE_TABLE rename column COLUMN_A to OLD_COLUMN_A;
alter table OOPSIE_TABLE rename column COLUMN_B to OLD_COLUMN_B;
-- Define COLUMN_A and COLUMN_B to have TIME ZONE support.
alter table OOPSIE_TABLE add (
COLUMN_A timestamp(6) with time zone,
COLUMN_B timestamp(6) with time zone
);
-- Populate the "new" columns with the adjusted version of the old data.
update OOPSIE_TABLE set
COLUMN_A = from_tz(OLD_COLUMN_A, 'America/New_York') at time zone 'UTC',
COLUMN_B = from_tz(OLD_COLUMN_B, 'America/New_York') at time zone 'UTC'
;
Run Code Online (Sandbox Code Playgroud)
| 归档时间: |
|
| 查看次数: |
8474 次 |
| 最近记录: |