Anmol Sundaram created HIVE-27004:
-------------------------------------
Summary: DateTimeFormatterBuilder cannot parse 'UTC+' in Java
versions higher than 8
Key: HIVE-27004
URL: https://issues.apache.org/jira/browse/HIVE-27004
Project: Hive
Issue Type: Bug
Components: Hive
Environment: Java version used
{{}}
{code:java}
openjdk version "11.0.16.1" 2022-08-12 LTS OpenJDK Runtime Environment
Corretto-11.0.16.9.1 (build 11.0.16.1+9-LTS) OpenJDK 64-Bit Server VM
Corretto-11.0.16.9.1 (build 11.0.16.1+9-LTS, mixed mode{code}
{{}}
Reporter: Anmol Sundaram
Assignee: Anmol Sundaram
Some of the unit tests related to _DateTimeFormatter_ were failing in Java
versions greater than 8 while working in Java 8.
Example of a failing Unit Test :
_org.apache.hadoop.hive.common.type.TestTimestampTZ#testComparision_
{code:java}
java.time.format.DateTimeParseException: Text '2017-04-14 18:00:00 UTC+08:00'
could not be parsed, unparsed text found at index 23 at
java.base/java.time.format.DateTimeFormatter.parseResolved0(DateTimeFormatter.java:2049)
at
java.base/java.time.format.DateTimeFormatter.parse(DateTimeFormatter.java:1874)
at
org.apache.hadoop.hive.common.type.TimestampTZUtil.parse(TimestampTZUtil.java:76)
at
org.apache.hadoop.hive.common.type.TimestampTZUtil.parse(TimestampTZUtil.java:64)
at
org.apache.hadoop.hive.common.type.TestTimestampTZ.testComparison(TestTimestampTZ.java:44)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native
Method){code}
The {{appendZoneText(TextStyle)}} method of
[DateTimeFormatteBuilder|https://github.com/apache/hive/blob/master/common/src/java/org/apache/hadoop/hive/common/type/TimestampTZUtil.java#L82]
is not able to parse the {{+}} symbol in {{2017-04-14 18:00:00 UTC+08:00}}
when running [the
test|https://github.com/apache/hive/blob/master/common/src/test/org/apache/hadoop/hive/common/type/TestTimestampTZ.java#L37]
in Java 11 , while it is working fine with Java 8.
According to the
[doc|https://developer.android.com/reference/java/time/format/DateTimeFormatterBuilder#appendZoneText(java.time.format.TextStyle)]
, {{appendZoneText(TextStyle)}} should be able to parse {{either the textual
zone name, the zone ID or the offset}} and UTC+08:00 should come under offset
as per the same
[doc|https://developer.android.com/reference/java/time/format/DateTimeFormatterBuilder#appendOffset(java.lang.String,%20java.lang.String)].
It seems, however, that "UTC" was explicitly removed [when parsed as ZoneText
for higher Java
versions|https://github.com/openjdk/jdk/commit/5c3a01591c5c945926636fdc9f164d60b5b4f29e?diff=unified#diff-5fcf976db1c06e8f44a8671356d7e34fdfbf5b057baa852e7c3e015c8797c889R4263]
As a workaround, we can use {{appendZoneOrOffsetId()}} rather than
{{appendZoneText().}}
This ensures the tests are passing for Java 8+ and based on my testing, I
didn’t see any regression of the change.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
