hanyuzheng7 commented on code in PR #25897:
URL: https://github.com/apache/flink/pull/25897#discussion_r1909349391
##########
flink-table/flink-table-planner/src/test/java/org/apache/flink/table/planner/functions/TimeFunctionsITCase.java:
##########
@@ -629,8 +645,9 @@ private Stream<TestSetSpec> floorTestCases() {
// Fractional seconds are lost
LocalTime.of(11, 22, 33),
LocalDate.of(1990, 10, 14),
- LocalDateTime.of(2020, 2, 29, 1, 56, 59,
987654321))
- .andDataTypes(TIME(), DATE(), TIMESTAMP())
+ LocalDateTime.of(2020, 2, 29, 1, 56, 59,
987654321),
Review Comment:
Yes, as a result, we don't use `f2` and value under `f2` for
`CEIL(CAST(LocalDateTime AS TIMESTAMP_LTZ()) TO
SECOND/MINUTE/MILLISECOND/MICROSECOND/NANOSECOND)` tests. All of them use `f3`.
I just let all tests such as `CAST(CEIL(CAST(f3 AS TIMESTAMP_LTZ(9)) TO
SECOND) AS STRING)` use `f3`
```
.testResult(
$("f3").cast(TIMESTAMP_LTZ(9))
.ceil(TimeIntervalUnit.SECOND)
.cast(STRING()),
"CAST(CEIL(CAST(f3 AS TIMESTAMP_LTZ(9)) TO
SECOND) AS STRING)",
LocalDateTime.of(2021, 9, 24, 9, 20, 51, 0)
.format(TIMESTAMP_FORMATTER),
STRING().nullable())
```
But the tests such as `CEIL(f2 TO DECADE)`
```
.testResult(
$("f2").ceil(TimeIntervalUnit.DECADE),
"CEIL(f2 TO DECADE)",
LocalDateTime.of(2030, 1, 1, 0, 0),
TIMESTAMP().nullable())
```
I do not change it, so we need `f2`, other tests still use it.
Tests using ` .format(TIMESTAMP_FORMATTER),` will use f3 now. Because we
change TIMESTAMP_FORMATTER from ` DateTimeFormatter.ofPattern("yyyy-MM-dd'
'HH:mm:ss.SSS");`
to `DateTimeFormatter.ofPattern("yyyy-MM-dd' 'HH:mm:ss.SSSSSSSSS");` and f3
precision is 9 to match the DateTimeFormatter.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
