Fix bug in discardNull (#5927)

* Fix bug in discardNull

* ci: auto code format

---------

Co-authored-by: CI Bot <[email protected]>

Author: andyfengHKU

src/common/vector/value_vector.cpp

@@ -58,25 +58,24 @@ uint32_t ValueVector::countNonNull() const {
 bool ValueVector::discardNull(ValueVector& vector) {
     if (vector.hasNoNullsGuarantee()) {
         return true;
+    }
+    auto selectedPos = 0u;
+    if (vector.state->getSelVector().isUnfiltered()) {
+        auto buffer = vector.state->getSelVectorUnsafe().getMutableBuffer();
+        for (auto i = 0u; i < vector.state->getSelVector().getSelSize(); i++) {
+            buffer[selectedPos] = i;
+            selectedPos += !vector.isNull(i);
+        }
+        vector.state->getSelVectorUnsafe().setToFiltered();
     } else {
-        auto selectedPos = 0u;
-        if (vector.state->getSelVector().isUnfiltered()) {
-            auto buffer = vector.state->getSelVectorUnsafe().getMutableBuffer();
-            for (auto i = 0u; i < vector.state->getSelVector().getSelSize(); i++) {
-                buffer[selectedPos] = i;
-                selectedPos += !vector.isNull(i);
-            }
-            vector.state->getSelVectorUnsafe().setToFiltered();
-        } else {
-            for (auto i = 0u; i < vector.state->getSelVector().getSelSize(); i++) {
-                auto pos = vector.state->getSelVector()[i];
-                vector.state->getSelVectorUnsafe()[i] = pos;
-                selectedPos += !vector.isNull(pos);
-            }
+        for (auto i = 0u; i < vector.state->getSelVector().getSelSize(); i++) {
+            auto pos = vector.state->getSelVector()[i];
+            vector.state->getSelVectorUnsafe()[selectedPos] = pos;
+            selectedPos += !vector.isNull(pos);
         }
-        vector.state->getSelVectorUnsafe().setSelSize(selectedPos);
-        return selectedPos > 0;
     }
+    vector.state->getSelVectorUnsafe().setSelSize(selectedPos);
+    return selectedPos > 0;
 }
 
 bool ValueVector::setNullFromBits(const uint64_t* srcNullEntries, uint64_t srcOffset,

src/processor/operator/hash_join/join_hash_table.cpp

@@ -136,14 +136,14 @@ void JoinHashTable::probe(const std::vector<ValueVector*>& keyVectors, ValueVect
         return;
     }
     hashSelVec.setSelSize(keyVectors[0]->state->getSelVector().getSelSize());
-    function::VectorHashFunction::computeHash(*keyVectors[0], keyVectors[0]->state->getSelVector(),
+    VectorHashFunction::computeHash(*keyVectors[0], keyVectors[0]->state->getSelVector(),
         hashVector, hashSelVec);
     for (auto i = 1u; i < keyVectors.size(); i++) {
         hashSelVec.setSelSize(keyVectors[i]->state->getSelVector().getSelSize());
-        function::VectorHashFunction::computeHash(*keyVectors[i],
-            keyVectors[i]->state->getSelVector(), *tmpHashResultVector, hashSelVec);
-        function::VectorHashFunction::combineHash(hashVector, hashSelVec, *tmpHashResultVector,
-            hashSelVec, hashVector, hashSelVec);
+        VectorHashFunction::computeHash(*keyVectors[i], keyVectors[i]->state->getSelVector(),
+            *tmpHashResultVector, hashSelVec);
+        VectorHashFunction::combineHash(hashVector, hashSelVec, *tmpHashResultVector, hashSelVec,
+            hashVector, hashSelVec);
     }
     for (auto i = 0u; i < hashSelVec.getSelSize(); i++) {
         KU_ASSERT(i < DEFAULT_VECTOR_CAPACITY);

test/test_files/issue/issue.test

@@ -592,3 +592,22 @@ Copy exception: Unable to find primary key value 1.
 -STATEMENT match (a:Test1), (b:Test2) create (a)-[e:Xyz]->(b);
 ---- error
 Binder exception: Cannot find a valid label in e that connects Test1 and Test2.
+
+-CASE 5893
+
+-STATEMENT create node table Account(id serial primary key, service string, name string);
+---- ok
+-STATEMENT create (a:Account {service: 'A', name: 'alice'});
+---- ok
+-STATEMENT create (a:Account {service: 'A', name: null});
+---- ok
+-STATEMENT create (a:Account {service: 'B', name: null});
+---- ok
+-STATEMENT create (a:Account {service: 'B', name: 'alice'});
+---- ok
+-STATEMENT match (a_account:Account {service: 'A'})
+    match (b_account:Account {service: 'B'})
+    where a_account.name = b_account.name
+    return a_account.name;
+---- 1
+alice