xuzifu666 commented on code in PR #4400:
URL: https://github.com/apache/calcite/pull/4400#discussion_r2110719414
##########
core/src/main/java/org/apache/calcite/rex/RexSimplify.java:
##########
@@ -2364,6 +2366,38 @@ private RexNode simplifyCeilFloor(RexCall e) {
ImmutableList.of(operand, e.getOperands().get(1)));
}
+ /** Simplify TRIM function by eliminating nested duplication.
+ *
+ * <p>Examples:
+ * <ul>
+ *
+ * <li>{@code trim(trim(' aa '))} returns {@code trim(' aa ')}
+ *
+ * <li>{@code trim(BOTH ' ' from trim(BOTH ' ' from ' aa '))}
+ * returns {@code trim(BOTH ' ' from ' aa ')}
+ *
+ * <li>{@code trim(LEADING 'a' from trim(BOTH ' ' from ' aa '))} does not
change
+ *
+ * </ul>
+ */
+ private RexNode simplifyTrim(RexCall e) {
+ if (e.getOperands().size() != 3) {
+ return e;
+ }
+
+ if (e.getOperands().get(2) instanceof RexCall) {
+ RexCall childNode = (RexCall) e.getOperands().get(2);
+ // only strings with the same trim method and deduplication will be
eliminated.
+ if (childNode.getKind() == SqlKind.TRIM
+ &&
simplify(e.operands.get(0)).equals(simplify(childNode.operands.get(0)))
Review Comment:
It has been specified as BOTH/LEADING/TAILING in JAVACC file.
If this place is changed to an expression or column, a syntax error will be
reported directly, so it may not be possible to add a test like this, but I
understand that there is no problem with the simplified call hereļ¼because it
must be a string literal. I wonder if you agree with my point of view?
@mihaibudiu
--
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]
