[HIVE-17923] 'cluster by' should not be needed for a bucketed table

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Duplicate
• Affects Version/s: 3.0.0
• Target Version/s: 3.0.0
• Assignee: Deepak Jaiswal Assign to me
• Reporter: Eugene Koifman
• Created: 27/Oct/17 15:58
• Updated: 07/Dec/17 20:26
• Resolved: 07/Dec/17 20:11

Description

given

CREATE TABLE over10k_orc_bucketed(t tinyint,
           si smallint,
           i int,
           b bigint,
           f float,
           d double,
           bo boolean,
           s string,
           ts timestamp,
           `dec` decimal(4,2),
           bin binary) CLUSTERED BY(si) INTO 4 BUCKETS STORED AS ORC;

insert into over10k_orc_bucketed select * from over10k

produces 1 data file (bucket 0).  It should produce 4 based on input data.

insert into over10k_orc_bucketed select * from over10k cluster by si

does the right thing.

acid_vectorization_original.q has the full script (HIVE-17458)

Issue Links

• blocks: HIVE-17458 VectorizedOrcAcidRowBatchReader doesn't handle 'original' files
• is duplicated by: HIVE-18157 Vectorization : Insert in bucketed table is broken with vectorization