[IMPALA-9921] Parse errors in ToSqlUtils.hiveNeedsQuotes should not be printed to impalad.ERROR

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: Impala 2.8.0, Impala 2.9.0, Impala 2.10.0, Impala 2.11.0, Impala 3.0, Impala 2.12.0, Impala 3.1.0, Impala 3.2.0, Impala 3.3.0, Impala 3.4.0
• Fix Version/s: Impala 4.0.0, Impala 3.4.1
• Component/s: Frontend
• Epic Color: ghx-label-9
• Assignee: Quanlong Huang
• Reporter: Quanlong Huang
• Created: 07/Jul/20 01:46
• Updated: 23/Mar/22 06:39
• Resolved: 08/Jul/20 05:36

Description

Impala planner will check whether an ident needs to be quoted in toSql outputs by using the HiveLexer (org.apache.hadoop.hive.ql.parse.HiveLexer). However, when there are parse errors from HiveLexer, the error messages will be printed to stderr which is redirected to impalad.ERROR. Actually, these error messages only means the ident needs to be quoted. They should have a lower log level and shouldn't be printed to impalad.ERROR.

A possible solution is to extend HiveLexer and override the displayRecognitionError() as HiveLexerX does. We can also consider using HiveLexerX directly.

Reproduction

impala-shell> select 'hello' as `你好`;

Some error messages are printed in impalad.ERROR:

line 1:0 no viable alternative at character '你'
line 1:1 no viable alternative at character '好'
line 1:0 no viable alternative at character '你'
line 1:1 no viable alternative at character '好'
line 1:0 no viable alternative at character '你'
line 1:1 no viable alternative at character '好'

Issue Links

• relates to: IMPALA-9641 Query hang when containing alias names as empty backticks
• links to: Code Review