[GitHub] [calcite] mihaibudiu commented on a diff in pull request #3430: [CALCITE-6008] ARRAY_AGG should returns ARRAY NULL when there's no input rows
mihaibudiu commented on code in PR #3430:
URL: https://github.com/apache/calcite/pull/3430#discussion_r183870
##
testkit/src/main/java/org/apache/calcite/test/SqlOperatorTest.java:
##
@@ -9314,11 +9314,12 @@ private static void
checkGroupConcatFuncFails(SqlOperatorFixture t) {
}
private static void checkArrayAggFunc(SqlOperatorFixture f) {
-f.setFor(SqlLibraryOperators.ARRAY_CONCAT_AGG, VM_FENNEL, VM_JAVA);
+f.setFor(SqlLibraryOperators.ARRAY_AGG, VM_FENNEL, VM_JAVA);
Review Comment:
This was just a suggestion, I don't think it is required necessarily for
this PR.
--
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: commits-unsubscr...@calcite.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
[GitHub] [calcite] mihaibudiu commented on a diff in pull request #3430: [CALCITE-6008] ARRAY_AGG should returns ARRAY NULL when there's no input rows
mihaibudiu commented on code in PR #3430:
URL: https://github.com/apache/calcite/pull/3430#discussion_r181033
##
testkit/src/main/java/org/apache/calcite/test/SqlOperatorTest.java:
##
@@ -9314,11 +9314,12 @@ private static void
checkGroupConcatFuncFails(SqlOperatorFixture t) {
}
private static void checkArrayAggFunc(SqlOperatorFixture f) {
-f.setFor(SqlLibraryOperators.ARRAY_CONCAT_AGG, VM_FENNEL, VM_JAVA);
+f.setFor(SqlLibraryOperators.ARRAY_AGG, VM_FENNEL, VM_JAVA);
Review Comment:
I have found that sometimes optimization rules produce results that differ
from the runtime results for the same expression. In particular
`CoreRules.PROJECT_REDUCE_EXPRESSIONS` will optimize constant expressions.
Currently there is no easy way to check that the optimization produces the
expected result. Does the SqlToRelConverterTest allow you to specify
optimizations?
I am working towards https://issues.apache.org/jira/browse/CALCITE-5891,
once that exists all the existing SqlOperatorTests will be automatically run in
both ways. (But it may take awhile, it has uncovered at least 50 bugs in
RelToSql and the optimizer.)
--
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: commits-unsubscr...@calcite.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
[GitHub] [calcite] mihaibudiu commented on a diff in pull request #3430: [CALCITE-6008] ARRAY_AGG should returns ARRAY NULL when there's no input rows
mihaibudiu commented on code in PR #3430:
URL: https://github.com/apache/calcite/pull/3430#discussion_r1327999598
##
testkit/src/main/java/org/apache/calcite/test/SqlOperatorTest.java:
##
@@ -9314,11 +9314,12 @@ private static void
checkGroupConcatFuncFails(SqlOperatorFixture t) {
}
private static void checkArrayAggFunc(SqlOperatorFixture f) {
-f.setFor(SqlLibraryOperators.ARRAY_CONCAT_AGG, VM_FENNEL, VM_JAVA);
+f.setFor(SqlLibraryOperators.ARRAY_AGG, VM_FENNEL, VM_JAVA);
Review Comment:
The optimization will replace the plan with a query returning constant null.
--
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: commits-unsubscr...@calcite.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
[GitHub] [calcite] mihaibudiu commented on a diff in pull request #3430: [CALCITE-6008] ARRAY_AGG should returns ARRAY NULL when there's no input rows
mihaibudiu commented on code in PR #3430:
URL: https://github.com/apache/calcite/pull/3430#discussion_r1327987695
##
testkit/src/main/java/org/apache/calcite/test/SqlOperatorTest.java:
##
@@ -9314,11 +9314,12 @@ private static void
checkGroupConcatFuncFails(SqlOperatorFixture t) {
}
private static void checkArrayAggFunc(SqlOperatorFixture f) {
-f.setFor(SqlLibraryOperators.ARRAY_CONCAT_AGG, VM_FENNEL, VM_JAVA);
+f.setFor(SqlLibraryOperators.ARRAY_AGG, VM_FENNEL, VM_JAVA);
Review Comment:
When fixing some similar bugs I have found that a RelOptRulesTest can often
exhibit the bug by producing a wrong plan.
Here is an example:
https://github.com/apache/calcite/pull/3289/files#diff-a430972be1169f680b7886b824720254fa1dabe37191db34bd2a63f10cf7a438
I don't know if it will work in your case as well.
--
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: commits-unsubscr...@calcite.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
