Created 01-19-2018 10:31 AM
Hello everyone ,
I have to update and maybe delete some rows in existing hive tables ( orc ) , unfortunately the update and delete operations are blocked by the datalake management team .
So , i would like to know if there is any way to do the update without direct acid operations
thx
Created 01-19-2018 01:10 PM
You can always create a new table based on a select that changes and removes the unwanted rows/data, and then rename the tables.
Created 01-19-2018 01:19 PM
thx berry , I'am doing this actually , but i was wondering if there is other "clean" solutions .
Created 01-19-2018 04:51 PM
Hi @Réda, unfortunately without the "transactional=true" there is no means to perform insert, update, or deletes on a Hive table. The good news is in HDP 3.0 due out mid-year, ACID will be the default for all Hive tables.
Obviously, I'd also question why the data lake team is blocking transactions. Many customer will allow for transactional tables at the very least in a staging area to handle CDC operations.