hadoop - Hive alter table change column name 将 'NULL' 赋予重命名的列

Question

我曾尝试将表中的现有列重命名为新列。但在名称更改后，新列只给我“NULL”值。

Parquet 中表的存储格式。

例如，

'user' is a column in 'Test' table of string data type. Inserted a sample record with value as 'John'.

Select user from Test;

Result : John

I have renamed 'user' to 'user_name' without changing any data type.

ALTER TABLE Test CHANGE user user_name String;

Select user_name from Test;

Result : NULL

请让我知道如何解决这个问题？

Whether MSCK Repair table command be of any use in this case?

Do I need to reload this table again to fix this issue?

问候， 阿达什KS

Answer 1

你可以做的是添加新字段，执行一次插入覆盖，然后删除旧字段。 像这样:

ALTER TABLE temp.Test ADD COLUMNS (user_new string) CASCADE;

insert overwrite table temp.Test
select 
      user_a,
      a,
      b,
      c,
      user_a as user_new
from temp.test;

ALTER TABLE temp.test  REPLACE COLUMNS(user_new string, a string, b string, c string );