[
https://issues.apache.org/jira/browse/OPENJPA-2863?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17317158#comment-17317158
]
Mark Struberg commented on OPENJPA-2863:
----------------------------------------
The constant {{LocalTime.MAX}} uses {{999999999}} for nanoseconds and your
column is only {{TIMESTAMP(6)}} thus the value gets rounded up. That part is
actually ok.
But I found another bug which is likely very related. Even the following code
will round over to the next day:
{noformat}
// zero out the nanos part and truncate to micros essentially:
awayUntil = LocalDate.of(2021, 4, 8).atTime(LocalTime.of(23, 59, 59, 666666000);
{noformat}
This is is the result of a bug we introduced a long time ago and also affects
our handling of {{java.sql.Timestamp}}. We round those values to MILLIS and not
to MICROS which is the default for Oracle.
I gonna fix that in {{OracleDictionary}}, for retaining the old 3 fraction
digits behaviour one can use the configuration in persistence.xml properties
section to use 1_000_000 nanosecond precision:
{noformat}
<property name="openjpa.jdbc.Dictionary" value="(datePrecision=1000000)"/>
{noformat}
> java.time.LocalDateTime in Oracle gets rounded to just 3 digits
> ---------------------------------------------------------------
>
> Key: OPENJPA-2863
> URL: https://issues.apache.org/jira/browse/OPENJPA-2863
> Project: OpenJPA
> Issue Type: Bug
> Components: jpa
> Affects Versions: 3.1.2
> Reporter: Karl Grosse
> Assignee: Mark Struberg
> Priority: Major
>
> Having a column like
>
> {code:java}
> @NotNull
> @Column(name = "awayUntil", nullable = false)
> private LocalDateTime awayUntil;
> {code}
> which we set to
> {code:java}
> awayUntil = LocalDate.of(2021, 4, 8).atTime(LocalTime.MAX);
> {code}
> gets somehow rounded to 2021-4-9 00:00:00,000000 after persisting into the
> Oracle DB.
> The type of the column in Oracle is Timestamp(6).
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
