Anthrino commented on code in PR #3677:
URL: https://github.com/apache/calcite/pull/3677#discussion_r1554345497
##########
core/src/main/java/org/apache/calcite/adapter/enumerable/RexToLixTranslator.java:
##########
@@ -469,34 +349,48 @@ private Expression getConvertExpression(
final SqlIntervalQualifier interval =
sourceType.getIntervalQualifier();
switch (sourceType.getSqlTypeName()) {
+ // If format string is supplied, return formatted date/time/timestamp
case DATE:
- return RexImpTable.optimize2(operand,
- Expressions.call(BuiltInMethod.UNIX_DATE_TO_STRING.method,
- operand));
+ return RexImpTable.optimize2(operand,
Expressions.isConstantNull(format)
+ ? Expressions.call(BuiltInMethod.UNIX_DATE_TO_STRING.method,
operand)
+ : Expressions.call(
+ Expressions.new_(
+ BuiltInMethod.FORMAT_DATE.method.getDeclaringClass()),
+ BuiltInMethod.FORMAT_DATE.method, format, operand));
case TIME:
- return RexImpTable.optimize2(operand,
- Expressions.call(BuiltInMethod.UNIX_TIME_TO_STRING.method,
- operand));
+ return RexImpTable.optimize2(operand,
Expressions.isConstantNull(format)
Review Comment:
Not exactly, we could only move the `RexImpTable.optimize2()` call outside
the switch, but the specific method called to evaluate the CAST depends on the
from and to cast data types, and with this change also if there is a `FORMAT`
present. So the number of lines would still look the same, but we could move
out one level of nesting with the optimize calls and return statements outside
the switch case.
Does that seem like a better pattern?
--
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]
