[DRILL-4688] String functions may produce incorrect result when input has multi-byte character

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Fix Version/s: 1.7.0
• Assignee: Unassigned
• Reporter: Jinfeng Ni
• Created: 19/May/16 17:20
• Updated: 16/Jun/16 03:39
• Resolved: 16/Jun/16 03:39

Description

As discussed in DRILL-4573, the patch of DRILL-4573 would cause regression of query correctness, when the input, encoded as utf-8, contains multi-byte characters.

For example,

select regexp_matches('München', 'München') res3 from (values(1));
+--------+
| res3 |
+--------+
| false |
+--------+

Here is the result before the patch of DRILL-4573.

select regexp_matches('München', 'München') res3 from (values(1));
+-------+
| res3 |
+-------+
| true |
+-------+

Once this issue has been resolved, QA would add functional testcases to cover the case of multip-byte characters, so that we will be able to catch such regression in the first place in the future.

Issue Links

• relates to: DRILL-4573 Zero copy LIKE, REGEXP_MATCHES, SUBSTR