[
https://issues.apache.org/jira/browse/FLINK-10489?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Flink Jira Bot updated FLINK-10489:
-----------------------------------
Labels: auto-deprioritized-major auto-unassigned stale-minor (was:
auto-deprioritized-major auto-unassigned)
I am the [Flink Jira Bot|https://github.com/apache/flink-jira-bot/] and I help
the community manage its development. I see this issues has been marked as
Minor but is unassigned and neither itself nor its Sub-Tasks have been updated
for 180 days. I have gone ahead and marked it "stale-minor". If this ticket is
still Minor, please either assign yourself or give an update. Afterwards,
please remove the label or in 7 days the issue will be deprioritized.
> Inconsistent window information for streams with EventTime characteristic
> -------------------------------------------------------------------------
>
> Key: FLINK-10489
> URL: https://issues.apache.org/jira/browse/FLINK-10489
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / API
> Environment: Flink v1.5.3
> Reporter: Alexis Sarda-Espinosa
> Priority: Minor
> Labels: auto-deprioritized-major, auto-unassigned, stale-minor
>
> Consider the following program.
> {code:java}
> void temp() throws Exception {
> StreamExecutionEnvironment execEnv =
> StreamExecutionEnvironment.getExecutionEnvironment();
> StreamTableEnvironment tableEnv =
> TableEnvironment.getTableEnvironment(execEnv);
> execEnv.setStreamTimeCharacteristic(TimeCharacteristic.EventTime);
> DataStream<Tuple3<Timestamp, String, Double>> source = execEnv
> .fromElements(
> new Tuple3<>(Timestamp.valueOf("2018-09-20 22:00:00"),
> "a", 1.3),
> new Tuple3<>(Timestamp.valueOf("2018-09-20 22:01:00"),
> "a", 2.1),
> new Tuple3<>(Timestamp.valueOf("2018-09-20 22:02:00"),
> "a", 3.0),
> new Tuple3<>(Timestamp.valueOf("2018-09-20 22:00:00"),
> "b", 2.2),
> new Tuple3<>(Timestamp.valueOf("2018-09-20 22:01:00"),
> "b", 1.8)
> )
> .assignTimestampsAndWatermarks(new WaterMarker());
> Table table = tableEnv.fromDataStream(source, "f0, f1, f2,
> rowtime.rowtime")
>
> .window(Slide.over("2.minutes").every("1.minute").on("rowtime").as("w"))
> .groupBy("f1, w")
> .select("f1, w.end - 1.minute, f2.sum");
> tableEnv.toAppendStream(table, Row.class).print();
> execEnv.execute();
> }
> public static class WaterMarker implements
> AssignerWithPunctuatedWatermarks<Tuple3<Timestamp, String, Double>> {
> @Nullable
> @Override
> public Watermark checkAndGetNextWatermark(Tuple3<Timestamp, String,
> Double> lastElement, long extractedTimestamp) {
> return new Watermark(extractedTimestamp);
> }
> @Override
> public long extractTimestamp(Tuple3<Timestamp, String, Double> element,
> long previousElementTimestamp) {
> return element.f0.getTime();
> }
> }
> {code}
> I am seeing an output like the following, where the hour is offset by 2 hours:
> {code:java}
> a,2018-09-20 20:00:00.0,1.3
> a,2018-09-20 20:01:00.0,3.4000000000000004
> a,2018-09-20 20:02:00.0,5.1
> a,2018-09-20 20:03:00.0,3.0
> {code}
> My local time zone is GMT+01:00, and indeed the input timestamps are parsed
> with this time zone, but even if the window-end is printed as GMT, it should
> show 21:0x:00. In my tests, where I actually collect the output and can see
> the actual {{java.sql.Timestamp}} instances, I can always see this extra 1
> hour offset.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
