fix

apache · Jan 2, 2024 · d7339d2 · d7339d2
1 parent a463f24
commit d7339d2
Show file tree

Hide file tree

Showing 5 changed files with 19 additions and 0 deletions.
diff --git a/cpp/core/jni/JniCommon.h b/cpp/core/jni/JniCommon.h
@@ -105,11 +105,13 @@ static inline jmethodID getStaticMethodIdOrError(JNIEnv* env, jclass thisClass,
 static inline void attachCurrentThreadAsDaemonOrThrow(JavaVM* vm, JNIEnv** out) {
   int getEnvStat = vm->GetEnv(reinterpret_cast<void**>(out), jniVersion);
   if (getEnvStat == JNI_EDETACHED) {
+    DLOG(INFO) << "JNIEnv was not attached to current thread.";
     // Reattach current thread to JVM
     getEnvStat = vm->AttachCurrentThreadAsDaemon(reinterpret_cast<void**>(out), NULL);
     if (getEnvStat != JNI_OK) {
       throw gluten::GlutenException("Failed to reattach current thread to JVM.");
     }
+    DLOG(INFO) << "Succeeded attaching current thread.";
     return;
   }
   if (getEnvStat != JNI_OK) {

diff --git a/cpp/core/shuffle/LocalPartitionWriter.cc b/cpp/core/shuffle/LocalPartitionWriter.cc
@@ -128,6 +128,7 @@ class FlushOnSpillEvictHandle final : public LocalPartitionWriter::LocalEvictHan
     ARROW_ASSIGN_OR_RAISE(auto start, os_->Tell());
     RETURN_NOT_OK(arrow::ipc::WriteIpcPayload(*payload, options_, os_.get(), &metadataLength));
     ARROW_ASSIGN_OR_RAISE(auto end, os_->Tell());
+    DLOG(INFO) << "Spilled partition " << partitionId << " file start: " << start << ", file end: " << end;
     spillInfo_->partitionSpillInfos.push_back({partitionId, end - start});
     return arrow::Status::OK();
   }

diff --git a/cpp/velox/benchmarks/common/OrcReaderIterator.h b/cpp/velox/benchmarks/common/OrcReaderIterator.h
@@ -61,6 +61,7 @@ class OrcStreamReaderIterator final : public OrcReaderIterator {
   std::shared_ptr<gluten::ColumnarBatch> next() override {
     auto startTime = std::chrono::steady_clock::now();
     GLUTEN_ASSIGN_OR_THROW(auto batch, recordBatchReader_->Next());
+    DLOG(INFO) << "OrcStreamReaderIterator get a batch, num rows: " << (batch ? batch->num_rows() : 0);
     collectBatchTime_ +=
         std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::steady_clock::now() - startTime).count();
     if (batch == nullptr) {
@@ -76,6 +77,12 @@ class OrcBufferedReaderIterator final : public OrcReaderIterator {
     createReader();
     collectBatches();
     iter_ = batches_.begin();
+    DLOG(INFO) << "OrcBufferedReaderIterator open file: " << path;
+    DLOG(INFO) << "Number of input batches: " << std::to_string(batches_.size());
+    if (iter_ != batches_.cend()) {
+      DLOG(INFO) << "columns: " << (*iter_)->num_columns();
+      DLOG(INFO) << "rows: " << (*iter_)->num_rows();
+    }
   }
 
   std::shared_ptr<gluten::ColumnarBatch> next() override {

diff --git a/cpp/velox/benchmarks/common/ParquetReaderIterator.h b/cpp/velox/benchmarks/common/ParquetReaderIterator.h
@@ -50,11 +50,13 @@ class ParquetStreamReaderIterator final : public ParquetReaderIterator {
  public:
   explicit ParquetStreamReaderIterator(const std::string& path) : ParquetReaderIterator(path) {
     createReader();
+    DLOG(INFO) << "ParquetStreamReaderIterator open file: " << path;
   }
 
   std::shared_ptr<gluten::ColumnarBatch> next() override {
     auto startTime = std::chrono::steady_clock::now();
     GLUTEN_ASSIGN_OR_THROW(auto batch, recordBatchReader_->Next());
+    DLOG(INFO) << "ParquetStreamReaderIterator get a batch, num rows: " << (batch ? batch->num_rows() : 0);
     collectBatchTime_ +=
         std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::steady_clock::now() - startTime).count();
     if (batch == nullptr) {
@@ -70,6 +72,12 @@ class ParquetBufferedReaderIterator final : public ParquetReaderIterator {
     createReader();
     collectBatches();
     iter_ = batches_.begin();
+    DLOG(INFO) << "ParquetBufferedReaderIterator open file: " << path;
+    DLOG(INFO) << "Number of input batches: " << std::to_string(batches_.size());
+    if (iter_ != batches_.cend()) {
+      DLOG(INFO) << "columns: " << (*iter_)->num_columns();
+      DLOG(INFO) << "rows: " << (*iter_)->num_rows();
+    }
   }
 
   std::shared_ptr<gluten::ColumnarBatch> next() override {

diff --git a/cpp/velox/compute/VeloxPlanConverter.cc b/cpp/velox/compute/VeloxPlanConverter.cc
@@ -211,6 +211,7 @@ std::shared_ptr<const facebook::velox::core::PlanNode> VeloxPlanConverter::toVel
     }
   }
   auto veloxPlan = substraitVeloxPlanConverter_.toVeloxPlan(substraitPlan);
+  DLOG(INFO) << "Plan Node: " << std::endl << veloxPlan->toString(true, true);
   return veloxPlan;
 }
-Original file line number
+Diff line change
@@ Expand Up @@
         }
       }
       auto veloxPlan = substraitVeloxPlanConverter_.toVeloxPlan(substraitPlan);
+      DLOG(INFO) << "Plan Node: " << std::endl << veloxPlan->toString(true, true);
       return veloxPlan;
     }
@@ Expand Down @@