[HIVE-21831] Stats should be reset correctly during load of a partitioned ACID table

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.1.0, 3.0.0, 3.1.1
• Fix Version/s: 4.0.0-alpha-1
• Component/s: Hive, Import/Export
• Labels: pull-request-available
• Assignee: David Lavati
• Reporter: David Lavati
• Created: 04/Jun/19 15:52
• Updated: 17/Nov/22 08:49
• Resolved: 12/Jun/19 07:32

Description

While running something similar to the following example, I noticed that an import of a partitioned ACID table using the ORC format fails to provide table statistics:

set hive.stats.autogather=true;
set hive.stats.column.autogather=true;
set hive.fetch.task.conversion=none;


set hive.support.concurrency=true;
set hive.default.fileformat.managed=ORC;
set hive.txn.manager=org.apache.hadoop.hive.ql.lockmgr.DbTxnManager;


create transactional table int_src (foo int, bar int);
insert into int_src select 1,1;


create transactional table int_exp(foo int) partitioned by (bar int);
insert into int_exp select * from int_src;
select count(*) from int_exp;


create transactional table int_imp(foo int) partitioned by (bar int);


EXPORT TABLE int_exp to '/tmp/expint';
IMPORT TABLE int_imp FROM '/tmp/expint';


select count(*) FROM int_imp;

The count returned 0 (opposed to 1, but even for 100k order of records it was 0) and correct statistics were only available after running compute statistics.

 

This was unique to ACID + partitioning + ORC, but this isn't the expected behavior.

Issue Links

• links to: GitHub Pull Request #659