[
https://issues.apache.org/jira/browse/ORC-87?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15984879#comment-15984879
]
ASF GitHub Bot commented on ORC-87:
-----------------------------------
Github user majetideepak commented on a diff in the pull request:
https://github.com/apache/orc/pull/110#discussion_r113462343
--- Diff: c++/src/Reader.cc ---
@@ -526,16 +526,25 @@ namespace orc {
if (metadata.get() == nullptr) {
throw std::logic_error("No stripe statistics in file");
}
+ auto currentStripeInfo =
footer->stripes(static_cast<int>(stripeIndex));
+ auto currentStripeFooter = getStripeFooter(currentStripeInfo,
*contents.get());
+ const Timezone& writerTZ =
+ currentStripeFooter.has_writertimezone() ?
+ getTimezoneByName(currentStripeFooter.writertimezone()) :
+ getLocalTimezone();
+ StatContext statContext;
+ statContext.correctStats = hasCorrectStatistics();
+ statContext.writerTimezone = &writerTZ;
return std::unique_ptr<Statistics>
(new StatisticsImpl(metadata->stripestats
- (static_cast<int>(stripeIndex)),
- hasCorrectStatistics()));
+ (static_cast<int>(stripeIndex)), statContext));
}
std::unique_ptr<Statistics> ReaderImpl::getStatistics() const {
+ StatContext statContext;
+ statContext.correctStats = hasCorrectStatistics();
--- End diff --
`getStatistics()` is returning the statistics in the file footer. As far as
I understood, there is no timezone information available in the file footer.
Looks like the timezone is stored for each stripe.
I am also curious about this design choice. I would imagine the file footer
to have the timezone information. Is it possible for different stripes to have
different timezone values?
> [C++] Handle missing timezone conversion for timestamp statistics
> -----------------------------------------------------------------
>
> Key: ORC-87
> URL: https://issues.apache.org/jira/browse/ORC-87
> Project: ORC
> Issue Type: Bug
> Components: C++
> Reporter: Deepak Majeti
> Assignee: Deepak Majeti
>
> The recent release adds timezone to timestamp values but does not add the
> conversion for timestamp statistics.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
