bvolpato commented on code in PR #39592:
URL: https://github.com/apache/beam/pull/39592#discussion_r3709588587
##########
sdks/java/extensions/sql/src/main/java/org/apache/beam/sdk/extensions/sql/impl/rel/BeamCalcRel.java:
##########
@@ -684,6 +699,12 @@ private static Expression toCalciteValue(
Expressions.multiply(dateValue,
Expressions.constant(MILLIS_PER_DAY)),
Expressions.divide(timeValue,
Expressions.constant(NANOS_PER_MILLISECOND)));
return nullOr(value, returnValue);
+ } else if
(org.apache.beam.sdk.schemas.logicaltypes.Timestamp.IDENTIFIER.equals(
+ identifier)) {
+ return nullOr(
+ value,
+ Expressions.call(
+ Expressions.convert_(value, java.time.Instant.class),
"toEpochMilli"));
Review Comment:
can contain sub-millisecond values, but this conversion goes through epoch
milliseconds. I reproduced this with : comes back as . Could we preserve
microsecond precision here, or reject values that Calcite cannot represent
without loss?
##########
sdks/java/extensions/sql/src/main/java/org/apache/beam/sdk/extensions/sql/impl/rel/BeamCalcRel.java:
##########
@@ -684,6 +699,12 @@ private static Expression toCalciteValue(
Expressions.multiply(dateValue,
Expressions.constant(MILLIS_PER_DAY)),
Expressions.divide(timeValue,
Expressions.constant(NANOS_PER_MILLISECOND)));
return nullOr(value, returnValue);
+ } else if
(org.apache.beam.sdk.schemas.logicaltypes.Timestamp.IDENTIFIER.equals(
+ identifier)) {
+ return nullOr(
+ value,
+ Expressions.call(
+ Expressions.convert_(value, java.time.Instant.class),
"toEpochMilli"));
Review Comment:
`Timestamp.MICROS` can contain sub-millisecond values, but this conversion
goes through epoch milliseconds. I reproduced this with `SELECT * FROM
PCOLLECTION`: `2025-07-31T20:17:40.123456Z` comes back as
`2025-07-31T20:17:40.123Z`. Could we preserve microsecond precision here, or
reject values that Calcite cannot represent without loss?
--
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]
