[impala] 01/03: IMPALA-11567: Fix left outer join if the right side is subquery with complex type

[jo...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

joemcdonnell pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/impala.git

commit 2e1ce445b24aba7d851475512a4c0378cbae0b4f
Author: Csaba Ringhofer <cs...@cloudera.com>
AuthorDate: Thu Sep 8 15:55:46 2022 +0200

    IMPALA-11567: Fix left outer join if the right side is subquery with complex type
    
    Non-matching rows from the left side will null out all slots from the
    right side in left outer joins. If the right side is a subquery, it
    is possible that some returned expressions will be non-NULL even if all
    slots are NULL (e.g. constants) - these expressions are wrapped as
    IF(TupleIsNull(tids), NULL, expr) to null them in the non-matching
    case.
    
    The logic above used to hit a precondition for complex types. We can
    safely ignore complex types for now, as currently the only possible
    expression that returns a complex type is SlotRef, which doesn't
    need to be wrapped. We will have to revisit this once functions are
    added that return complex types.
    
    Testing:
    - added a regression  test and ran it
    
    Change-Id: Iaa8991cd4448d5c7ef7f44f73ee07e2a2b6f37ce
    Reviewed-on: http://gerrit.cloudera.org:8080/18954
    Reviewed-by: Impala Public Jenkins <im...@cloudera.com>
    Tested-by: Impala Public Jenkins <im...@cloudera.com>
---
 .../org/apache/impala/analysis/TupleIsNullPredicate.java   | 10 ++++++++--
 .../queries/QueryTest/nested-array-in-select-list.test     | 14 ++++++++++++++
 2 files changed, 22 insertions(+), 2 deletions(-)

diff --git a/fe/src/main/java/org/apache/impala/analysis/TupleIsNullPredicate.java b/fe/src/main/java/org/apache/impala/analysis/TupleIsNullPredicate.java
index da78e6dc3..92685b8ba 100644
--- a/fe/src/main/java/org/apache/impala/analysis/TupleIsNullPredicate.java
+++ b/fe/src/main/java/org/apache/impala/analysis/TupleIsNullPredicate.java
@@ -162,8 +162,14 @@ public class TupleIsNullPredicate extends Predicate {
   public static boolean requiresNullWrapping(Expr expr, Analyzer analyzer)
       throws InternalException {
     Preconditions.checkNotNull(expr);
-    Preconditions.checkState(!expr.getType().isComplexType(),
-        "Should not evaluate on complex type: " + expr.debugString());
+
+    if (expr.getType().isComplexType()) {
+      // Currently the only Expr supported for complex types is SlotRef, which does not
+      // need NULL wrapping.
+      Preconditions.checkState(expr instanceof SlotRef);
+      return false;
+    }
+
     // If the expr is already wrapped in an IF(TupleIsNull(), NULL, expr)
     // then it must definitely be wrapped again at this level.
     // Do not try to execute expr because a TupleIsNullPredicate is not constant.
diff --git a/testdata/workloads/functional-query/queries/QueryTest/nested-array-in-select-list.test b/testdata/workloads/functional-query/queries/QueryTest/nested-array-in-select-list.test
index 14c1c54d1..f5ef5612d 100644
--- a/testdata/workloads/functional-query/queries/QueryTest/nested-array-in-select-list.test
+++ b/testdata/workloads/functional-query/queries/QueryTest/nested-array-in-select-list.test
@@ -391,4 +391,18 @@ select id, arr_int_1d, arr_int_2d, arr_int_3d, arr_string_1d, arr_string_2d, arr
 1,'[1,2,NULL]','[[1,2,NULL],[3]]','[[[1,2,NULL],[3]],[[4]]]','["1","2",NULL]','[["1","2",NULL],["3"]]','[[["1","2",NULL],["3"]],[["4"]]]'
 ---- TYPES
 INT,STRING,STRING,STRING,STRING,STRING,STRING
+====
+---- QUERY
+# Regression test for:
+#   IMPALA-11567: "Error in left outer join if the right side is a subquery with complex types"
+select a.id, b.arr_int_1d, b.arr_int_2d
+from alltypestiny a left outer join
+  (select id, arr_int_1d, arr_int_2d from collection_tbl) b
+on a.id = b.id where a.id < 3;
+---- RESULTS
+0,'NULL','NULL'
+1,'[1,2,NULL]','[[1,2,NULL],[3]]'
+2,'NULL','NULL'
+---- TYPES
+INT,STRING,STRING
 =====