Hi Igor,
Thanks for reply.
Your workaround need to touch Hive metastore but it is impossible for us so
far.
May I know is there a plan to fix this issue from Drill side as we
are eager to get it work.
Thanks
Jerry
Igor Guzenko 于2019年9月4日周三 下午4:45写道:
> Hello Jerry,
>
> Thank you very much for
Hello Jerry,
Thank you very much for detailed steps. The issue is pretty complex, when
Drill reads meta about partition, somehow linked
storage descriptor doesn't have the newly added column. I've created
DRILL-7365 [1]
which at the end of description contains one possible workaround.
[1]
Sorry, resend in pure text for cases mentioned before:
My case is as below:
1) create table hive_bucketed2 (emp_id int, first_name string) PARTITIONED
BY (`col_year_month` string) clustered by (emp_id) into 4 buckets stored as
orc tblproperties ('transactional'='true');
2) insert into
Hi Igor,
Thanks for your reply.
For you case Drill works fine.
But if the table created with Partition, will hit the issue.
My case is as below:
1) *create* *table *hive_bucketed2 (emp_id *int*, first_name *string*)
PARTITIONED *BY* (`col_year_month` *string*) clustered *by* (emp_id) *into*
4
Hello Jerry,
Could you please describe your use-case in details. I've tried following
steps and Drill reads data for new column just fine.
1) Create table
create table hive_bucketed(emp_id int, first_name string) clustered by
(emp_id) into 4 buckets stored as orc tblproperties
Hi,
We are using Drill v1.16.0 to query transactional (ACID) Hive bucketed ORC
table, everything is okay but when querying the columns added by HiveQL “alert
table … add columns …”, Drill always returns the query result as NULL for such
columns, HiveQL can query with expected results, can