[HIVE-25420] Ignore time type column in Iceberg testing for vectorized runs

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 4.0.0-alpha-1
• Labels: pull-request-available
• Assignee: Ádám Szita
• Reporter: Ádám Szita
• Created: 04/Aug/21 09:19
• Updated: 17/Nov/22 08:48
• Resolved: 05/Aug/21 09:50

Description

Time is a valid type in Iceberg but not in Hive. In Hive it is represented as a string type column, while (at least if ORC is used as underlying file format) long type is written out to data files.

This requires translation two times: long@ORC -> LocalDate@Iceberg -> toString()@Hive and it works well for non vectorized reads, but when vectorization is turned on, we will get:

org.apache.hadoop.hive.ql.exec.vector.BytesColumnVector cannot be cast to org.apache.hadoop.hive.ql.exec.vector.LongColumnVector 

Thus for now, time type is not supported with vectorization, and the relevant test cases should be ignored in such test configs.

Issue Links

• links to: GitHub Pull Request #2557