From 8db85d653391daf0acc21f2b01cb901e471bafc9 Mon Sep 17 00:00:00 2001
From: zhejiangxiaomai <zhenhui.zhao@intel.com>
Date: Thu, 30 Mar 2023 00:26:35 +0800
Subject: [PATCH] rebase to 2023/3/29

---
 velox/exec/Task.cpp                                       | 8 ++++----
 .../registration/DateTimeFunctionsRegistration.cpp        | 4 ----
 velox/functions/sparksql/MightContain.cpp                 | 4 ++--
 velox/functions/sparksql/Register.cpp                     | 3 +++
 .../sparksql/aggregates/BloomFilterAggAggregate.cpp       | 3 ++-
 5 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/velox/exec/Task.cpp b/velox/exec/Task.cpp
index a2d7041b0696..2a59bdd8b033 100644
--- a/velox/exec/Task.cpp
+++ b/velox/exec/Task.cpp
@@ -821,10 +821,10 @@ void Task::removeDriver(std::shared_ptr<Task> self, Driver* driver) {
     }
 
     if (self->numFinishedDrivers_ == self->numTotalDrivers_) {
-      LOG(INFO) << "All drivers (" << self->numFinishedDrivers_
+      /*LOG(INFO) << "All drivers (" << self->numFinishedDrivers_
                 << ") finished for task " << self->taskId()
                 << " after running for " << self->timeSinceStartMsLocked()
-                << " ms.";
+                << " ms.";*/
     }
   }
 
@@ -1470,9 +1470,9 @@ ContinueFuture Task::terminate(TaskState terminalState) {
       }
     }
 
-    LOG(INFO) << "Terminating task " << taskId() << " with state "
+    /*LOG(INFO) << "Terminating task " << taskId() << " with state "
               << taskStateString(state_) << " after running for "
-              << timeSinceStartMsLocked() << " ms.";
+              << timeSinceStartMsLocked() << " ms.";*/
 
     activateTaskCompletionNotifier(completionNotifier);
 
diff --git a/velox/functions/prestosql/registration/DateTimeFunctionsRegistration.cpp b/velox/functions/prestosql/registration/DateTimeFunctionsRegistration.cpp
index 4d3fb44d35cf..6cc1fdbf3cf6 100644
--- a/velox/functions/prestosql/registration/DateTimeFunctionsRegistration.cpp
+++ b/velox/functions/prestosql/registration/DateTimeFunctionsRegistration.cpp
@@ -32,10 +32,6 @@ void registerSimpleFunctions(const std::string& prefix) {
   registerFunction<YearFunction, int32_t, Date>({prefix + "year"});
   registerFunction<YearFunction, int32_t, TimestampWithTimezone>(
       {prefix + "year"});
-  registerFunction<YearFunction, int64_t, Timestamp>({prefix + "year"});
-  registerFunction<YearFunction, int64_t, Date>({prefix + "year"});
-  registerFunction<YearFunction, int64_t, TimestampWithTimezone>(
-      {prefix + "year"});
   registerFunction<WeekFunction, int64_t, Timestamp>(
       {prefix + "week", prefix + "week_of_year"});
   registerFunction<WeekFunction, int64_t, Date>(
diff --git a/velox/functions/sparksql/MightContain.cpp b/velox/functions/sparksql/MightContain.cpp
index 8fa26b091f90..f105b547cfd0 100644
--- a/velox/functions/sparksql/MightContain.cpp
+++ b/velox/functions/sparksql/MightContain.cpp
@@ -47,7 +47,7 @@ class BloomFilterMightContainFunction final : public exec::VectorFunction {
     HashStringAllocator allocator{context.pool()};
     if (serialized->isConstantMapping()) {
       BloomFilter output{StlAllocator<uint64_t>(&allocator)};
-      output.merge(serialized->valueAt<StringView>(0).data());
+      output.merge(serialized->valueAt<StringView>(0).str().c_str());
       rows.applyToSelected([&](int row) {
         auto contain = output.mayContain(
             folly::hasher<int64_t>()(value->valueAt<int64_t>(row)));
@@ -58,7 +58,7 @@ class BloomFilterMightContainFunction final : public exec::VectorFunction {
 
     rows.applyToSelected([&](int row) {
       BloomFilter output{StlAllocator<uint64_t>(&allocator)};
-      output.merge(serialized->valueAt<StringView>(0).data());
+      output.merge(serialized->valueAt<StringView>(0).str().c_str());
       auto contain = output.mayContain(
           folly::hasher<int64_t>()(value->valueAt<int64_t>(row)));
       result.set(row, contain);
diff --git a/velox/functions/sparksql/Register.cpp b/velox/functions/sparksql/Register.cpp
index 52c0dfa3ad20..d9eb29e59aef 100644
--- a/velox/functions/sparksql/Register.cpp
+++ b/velox/functions/sparksql/Register.cpp
@@ -55,6 +55,7 @@ static void workAroundRegistrationMacro(const std::string& prefix) {
       udf_array_intersect, prefix + "array_intersect");
   // This is the semantics of spark.sql.ansi.enabled = false.
   VELOX_REGISTER_VECTOR_FUNCTION(udf_element_at, prefix + "element_at");
+  VELOX_REGISTER_VECTOR_FUNCTION(udf_concat_row, prefix + "named_struct");
   VELOX_REGISTER_VECTOR_FUNCTION(
       udf_map_allow_duplicates, prefix + "map_from_arrays");
   // String functions.
@@ -117,6 +118,8 @@ void registerFunctions(const std::string& prefix) {
       prefix + "greatest", greatestSignatures(), makeGreatest);
   exec::registerStatefulVectorFunction(
       prefix + "hash", hashSignatures(), makeHash);
+  exec::registerStatefulVectorFunction(
+      prefix + "murmur3hash", hashSignatures(), makeHash);
   exec::registerStatefulVectorFunction(
       prefix + "xxhash64", xxhash64Signatures(), makeXxHash64);
   VELOX_REGISTER_VECTOR_FUNCTION(udf_map, prefix + "map");
diff --git a/velox/functions/sparksql/aggregates/BloomFilterAggAggregate.cpp b/velox/functions/sparksql/aggregates/BloomFilterAggAggregate.cpp
index 00d909972958..beeb22d65e1c 100644
--- a/velox/functions/sparksql/aggregates/BloomFilterAggAggregate.cpp
+++ b/velox/functions/sparksql/aggregates/BloomFilterAggAggregate.cpp
@@ -165,7 +165,8 @@ class BloomFilterAggAggregate : public exec::Aggregate {
         continue;
       }
       if (StringView::isInline(size)) {
-        StringView serialized(size);
+        char buffer[StringView::kInlineSize];
+        StringView serialized = StringView(buffer, size);
         accumulator->serialize(serialized);
         flatResult->setNoCopy(i, serialized);
       } else {