[IMPALA-5961] Test data for TPC-DS schema contains a non-Unicode character

• Type: Task
• Status: Reopened
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: Impala 2.10.0
• Component/s: Infrastructure
• Labels: TPCDS
• Epic Color: ghx-label-5
• Assignee: Unassigned
• Reporter: Tim Wood
• Created: 20/Sep/17 00:09
• Updated: 15/Dec/20 20:03

Description

The customer table contains rows whose c_birth_country values contain character 0xd4 (o-circumflex) in an illegal position for Unicode. This causes tpcds-q30 to fail. Either tests need to change to accommodate the different character set, or the test data should change to contain the proper Unicode character.

To reproduce, build a mini-cluster and load with test data (./buildall.sh -testdata ...), then run the query from the attached file. Find the affected rows with:
SELECT * FROM customer WHERE c_birth_country LIKE '%IVOIRE%';

Issue Links

• is required by: IMPALA-5376 Add all missing TPC-DS queries to functional tests