dawidwys commented on code in PR #22775:
URL: https://github.com/apache/flink/pull/22775#discussion_r1627434353
##########
flink-table/flink-table-planner/src/test/java/org/apache/flink/table/planner/functions/CastFunctionITCase.java:
##########
@@ -763,6 +763,68 @@ private static List<TestSetSpec> allTypesBasic() {
// ROW
// RAW
.build(),
+ CastTestSpecBuilder.testCastTo(TIME(2))
+ .fromCase(TIME(), null, null)
+ .failRuntime(CHAR(3), "foo", DateTimeException.class)
+ .failRuntime(VARCHAR(5), "Flink",
DateTimeException.class)
+ .failRuntime(STRING(), "Flink",
DateTimeException.class)
+ .fromCase(STRING(), "123", LocalTime.of(23, 0, 0))
+ .fromCase(STRING(), "123:45", LocalTime.of(23, 45, 0))
+ .failRuntime(STRING(), "2021-09-27",
DateTimeException.class)
+ .failRuntime(STRING(), "2021-09-27 12:34:56",
DateTimeException.class)
Review Comment:
I am wondering if we should duplicate those. Internally, there is no
difference if we cast those cases to `TIME(2)` or `TIME(0)`.
Personally, I'd be leaning to leave only the cases where we truncate the
precision.
##########
flink-table/flink-table-planner/src/test/java/org/apache/flink/table/planner/functions/CastFunctionITCase.java:
##########
@@ -763,6 +763,68 @@ private static List<TestSetSpec> allTypesBasic() {
// ROW
// RAW
.build(),
+ CastTestSpecBuilder.testCastTo(TIME(2))
+ .fromCase(TIME(), null, null)
+ .failRuntime(CHAR(3), "foo", DateTimeException.class)
+ .failRuntime(VARCHAR(5), "Flink",
DateTimeException.class)
+ .failRuntime(STRING(), "Flink",
DateTimeException.class)
+ .fromCase(STRING(), "123", LocalTime.of(23, 0, 0))
+ .fromCase(STRING(), "123:45", LocalTime.of(23, 45, 0))
+ .failRuntime(STRING(), "2021-09-27",
DateTimeException.class)
+ .failRuntime(STRING(), "2021-09-27 12:34:56",
DateTimeException.class)
+ .fromCase(
+ STRING(),
+ "12:34:56.123456789",
+ LocalTime.of(12, 34, 56, 120_000_000))
+ .failRuntime(
+ STRING(), "2021-09-27 12:34:56.123456789",
DateTimeException.class)
+ // Not supported - no fix
+ .failValidation(BOOLEAN(), true)
+ .failTableApiValidation(BINARY(2), DEFAULT_BINARY)
+ .failTableApiValidation(VARBINARY(5),
DEFAULT_VARBINARY)
+ .failTableApiValidation(BYTES(), DEFAULT_BYTES)
+ .failValidation(DECIMAL(5, 3), 12.345)
+ .failValidation(TINYINT(), DEFAULT_POSITIVE_TINY_INT)
+ .failValidation(SMALLINT(), DEFAULT_POSITIVE_SMALL_INT)
+ .failValidation(INT(), DEFAULT_POSITIVE_INT)
+ .failValidation(BIGINT(), DEFAULT_POSITIVE_BIGINT)
+ .failValidation(FLOAT(), DEFAULT_POSITIVE_FLOAT)
+ .failValidation(DOUBLE(), DEFAULT_POSITIVE_DOUBLE)
+ .failValidation(DATE(), DEFAULT_DATE)
+ //
+ .fromCase(TIME(5), DEFAULT_TIME, LocalTime.of(12, 34,
56, 120_000_000))
Review Comment:
Shouldn't this fail? Can we handle `TIME` with precision of `5`?
##########
flink-table/flink-table-planner/src/test/java/org/apache/flink/table/planner/functions/CastFunctionITCase.java:
##########
@@ -1038,9 +1112,9 @@ private static List<TestSetSpec> toStringCasts() {
DEFAULT_NEGATIVE_DOUBLE,
String.valueOf(DEFAULT_NEGATIVE_DOUBLE))
.fromCase(DATE(), DEFAULT_DATE, "2021-09-24")
- // https://issues.apache.org/jira/browse/FLINK-17224
Currently, fractional
- // seconds are lost
- .fromCase(TIME(5), DEFAULT_TIME, "12:34:56")
+ .fromCase(TIME(5), DEFAULT_TIME, "12:34:56.123")
Review Comment:
This shows, sth is wrong with `TIME(5)`.
--
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]
