vector: Fix join by array meet error "no supertype for types Array(Fl…

…oat32), Nullable(Array(Float32))" (pingcap#9490) ref pingcap#9032 Co-authored-by: ti-chi-bot[bot] <108142056+ti-chi-bot[bot]@users.noreply.github.com>
JaySon-Huang · Sep 29, 2024 · 1f30845 · 1f30845
1 parent 698a9ef
commit 1f30845
Show file tree

Hide file tree

Showing 5 changed files with 73 additions and 66 deletions.
diff --git a/dbms/src/DataTypes/getLeastSupertype.cpp b/dbms/src/DataTypes/getLeastSupertype.cpp
@@ -100,36 +100,6 @@ DataTypePtr getLeastSupertype(const DataTypes & types)
             return getLeastSupertype(non_nothing_types);
     }
 
-    /// For Arrays
-    {
-        bool have_array = false;
-        bool all_arrays = true;
-
-        DataTypes nested_types;
-        nested_types.reserve(types.size());
-
-        for (const auto & type : types)
-        {
-            if (const DataTypeArray * type_array = typeid_cast<const DataTypeArray *>(type.get()))
-            {
-                have_array = true;
-                nested_types.emplace_back(type_array->getNestedType());
-            }
-            else
-                all_arrays = false;
-        }
-
-        if (have_array)
-        {
-            if (!all_arrays)
-                throw Exception(
-                    getExceptionMessagePrefix(types) + " because some of them are Array and some of them are not",
-                    ErrorCodes::NO_COMMON_TYPE);
-
-            return std::make_shared<DataTypeArray>(getLeastSupertype(nested_types));
-        }
-    }
-
     /// For tuples
     {
         bool have_tuple = false;
@@ -204,6 +174,36 @@ DataTypePtr getLeastSupertype(const DataTypes & types)
         }
     }
 
+    /// For Arrays, canBeInsideNullable = true, should check it after handling Nullable
+    {
+        bool have_array = false;
+        bool all_arrays = true;
+
+        DataTypes nested_types;
+        nested_types.reserve(types.size());
+
+        for (const auto & type : types)
+        {
+            if (const DataTypeArray * type_array = typeid_cast<const DataTypeArray *>(type.get()))
+            {
+                have_array = true;
+                nested_types.emplace_back(type_array->getNestedType());
+            }
+            else
+                all_arrays = false;
+        }
+
+        if (have_array)
+        {
+            if (!all_arrays)
+                throw Exception(
+                    getExceptionMessagePrefix(types) + " because some of them are Array and some of them are not",
+                    ErrorCodes::NO_COMMON_TYPE);
+
+            return std::make_shared<DataTypeArray>(getLeastSupertype(nested_types));
+        }
+    }
+
     /// Non-recursive rules
 
     std::unordered_set<TypeIndex> type_ids;

diff --git a/dbms/src/DataTypes/getMostSubtype.cpp b/dbms/src/DataTypes/getMostSubtype.cpp
@@ -105,34 +105,6 @@ DataTypePtr getMostSubtype(const DataTypes & types, bool throw_if_result_is_noth
                 return get_nothing_or_throw(" because some of them are Nothing");
     }
 
-    /// For Arrays
-    {
-        bool have_array = false;
-        bool all_arrays = true;
-
-        DataTypes nested_types;
-        nested_types.reserve(types.size());
-
-        for (const auto & type : types)
-        {
-            if (const auto * const type_array = typeid_cast<const DataTypeArray *>(type.get()))
-            {
-                have_array = true;
-                nested_types.emplace_back(type_array->getNestedType());
-            }
-            else
-                all_arrays = false;
-        }
-
-        if (have_array)
-        {
-            if (!all_arrays)
-                return get_nothing_or_throw(" because some of them are Array and some of them are not");
-
-            return std::make_shared<DataTypeArray>(getMostSubtype(nested_types, false, force_support_conversion));
-        }
-    }
-
     /// For tuples
     {
         bool have_tuple = false;
@@ -210,6 +182,34 @@ DataTypePtr getMostSubtype(const DataTypes & types, bool throw_if_result_is_noth
         }
     }
 
+    /// For Arrays, canBeInsideNullable = true, should check it after handling Nullable
+    {
+        bool have_array = false;
+        bool all_arrays = true;
+
+        DataTypes nested_types;
+        nested_types.reserve(types.size());
+
+        for (const auto & type : types)
+        {
+            if (const auto * const type_array = typeid_cast<const DataTypeArray *>(type.get()))
+            {
+                have_array = true;
+                nested_types.emplace_back(type_array->getNestedType());
+            }
+            else
+                all_arrays = false;
+        }
+
+        if (have_array)
+        {
+            if (!all_arrays)
+                return get_nothing_or_throw(" because some of them are Array and some of them are not");
+
+            return std::make_shared<DataTypeArray>(getMostSubtype(nested_types, false, force_support_conversion));
+        }
+    }
+
     /// Non-recursive rules
 
     /// For String and FixedString, the common type is FixedString.

diff --git a/dbms/src/DataTypes/tests/gtest_data_type_get_common_type.cpp b/dbms/src/DataTypes/tests/gtest_data_type_get_common_type.cpp
@@ -150,10 +150,14 @@ try
                     ->equals(*typeFromString("Array(Array(UInt8))")));
     ASSERT_TRUE(getLeastSupertype(typesFromString("Array(Array(UInt8)) Array(Array(Int8))"))
                     ->equals(*typeFromString("Array(Array(Int16))")));
-    ASSERT_TRUE(
-        getLeastSupertype(typesFromString("Array(Date) Array(DateTime)"))->equals(*typeFromString("Array(DateTime)")));
+    ASSERT_TRUE(getLeastSupertype(typesFromString("Array(Date) Array(DateTime)")) //
+                    ->equals(*typeFromString("Array(DateTime)")));
     ASSERT_TRUE(getLeastSupertype(typesFromString("Array(String) Array(FixedString(32))"))
                     ->equals(*typeFromString("Array(String)")));
+    ASSERT_TRUE(getLeastSupertype(typesFromString("Array(Float32) Array(Float32)")) //
+                    ->equals(*typeFromString("Array(Float32)")));
+    ASSERT_TRUE(getLeastSupertype(typesFromString("Array(Float32) Nullable(Array(Float32))")) //
+                    ->equals(*typeFromString("Nullable(Array(Float32))")));
 
     ASSERT_TRUE(
         getLeastSupertype(typesFromString("Nullable(Nothing) Nothing"))->equals(*typeFromString("Nullable(Nothing)")));
@@ -214,11 +218,16 @@ try
                     ->equals(*typeFromString("Array(Array(UInt8))")));
     ASSERT_TRUE(getMostSubtype(typesFromString("Array(Array(UInt8)) Array(Array(Int8))"))
                     ->equals(*typeFromString("Array(Array(UInt8))")));
-    ASSERT_TRUE(getMostSubtype(typesFromString("Array(Date) Array(DateTime)"))->equals(*typeFromString("Array(Date)")));
+    ASSERT_TRUE(getMostSubtype(typesFromString("Array(Date) Array(DateTime)")) //
+                    ->equals(*typeFromString("Array(Date)")));
     ASSERT_TRUE(getMostSubtype(typesFromString("Array(String) Array(FixedString(32))"))
                     ->equals(*typeFromString("Array(FixedString(32))")));
     ASSERT_TRUE(getMostSubtype(typesFromString("Array(String) Array(FixedString(32))"))
                     ->equals(*typeFromString("Array(FixedString(32))")));
+    ASSERT_TRUE(getMostSubtype(typesFromString("Array(Float32) Array(Float32)")) //
+                    ->equals(*typeFromString("Array(Float32)")));
+    ASSERT_TRUE(getMostSubtype(typesFromString("Array(Float32) Nullable(Array(Float32))")) //
+                    ->equals(*typeFromString("Array(Float32)")));
 
     ASSERT_TRUE(getMostSubtype(typesFromString("Nullable(Nothing) Nothing"))->equals(*typeFromString("Nothing")));
     ASSERT_TRUE(getMostSubtype(typesFromString("Nullable(UInt8) Int8"))->equals(*typeFromString("UInt8")));

diff --git a/dbms/src/Flash/Mpp/MPPTaskScheduleEntry.cpp b/dbms/src/Flash/Mpp/MPPTaskScheduleEntry.cpp
@@ -42,11 +42,7 @@ bool MPPTaskScheduleEntry::schedule(ScheduleState state)
     if (schedule_state == ScheduleState::WAITING)
     {
         auto log_level = state == ScheduleState::SCHEDULED ? Poco::Message::PRIO_DEBUG : Poco::Message::PRIO_WARNING;
-        LOG_IMPL(
-            log,
-            log_level,
-            "task is {}.",
-            state == ScheduleState::SCHEDULED ? "scheduled" : " failed to schedule");
+        LOG_IMPL(log, log_level, "task is {}.", state == ScheduleState::SCHEDULED ? "scheduled" : "failed to schedule");
         schedule_state = state;
         schedule_cv.notify_one();
         return true;

diff --git a/dbms/src/Storages/DeltaMerge/ColumnStat.h b/dbms/src/Storages/DeltaMerge/ColumnStat.h
@@ -178,7 +178,9 @@ readText(ColumnStats & column_sats, DMFileFormat::Version ver, ReadBuffer & buf)
                 .serialized_bytes = serialized_bytes,
                 // ... here ignore some fields with default initializers
                 .vector_index = {},
+#ifndef NDEBUG
                 .additional_data_for_test = {},
+#endif
             });
     }
 }