hive 列名称更改而不会丢失分区 avro 表中的数据

Question

i have a avro hive table partitioned by date.

acct_no <string> acct_cd <int>  date <date>
acct01           01             2022-10-31
acct02           02             2022-10-31 
acct03           03             2022-10-31  

i want to rename acct_cd to account_code

i used command

alter table <table_name> change acct_cd account_code int;

it changed the table attribute but data lost:

acct_no <string> account_code <int>  date <date>
acct01         null                  2022-10-31
acct02         null                  2022-10-31 
acct03         null                  2022-10-31 

How to avoid losing data and change the column name?

Answer 1

As per hive, hive , change a column name just changes hive metadata, underlying data/structure remains as is. Which means, you have to take care of the data after alter.

You have two options -

1. Create a backup table. Drop original table. Crete a new table with new column name and insert into the new table from backup table.
2. Create a backup table. Rename the column in original table. Use insert overwrite table command to fillup original table from backup table.