minor changes

kuzudb · Jun 21, 2023 · dc8745b · dc8745b
1 parent 66edb07
commit dc8745b
Show file tree

Hide file tree

Showing 5 changed files with 9 additions and 9 deletions.
diff --git a/src/include/storage/copier/npy_reader.h b/src/include/storage/copier/npy_reader.h
@@ -51,6 +51,7 @@ class NpyReader {
     size_t dataOffset;
     std::vector<size_t> shape;
     common::LogicalTypeID type;
+    static inline const std::string defaultFieldName = "NPY_FIELD";
 };
 
 } // namespace storage

diff --git a/src/include/storage/in_mem_storage_structure/in_mem_column_chunk.h b/src/include/storage/in_mem_storage_structure/in_mem_column_chunk.h
@@ -45,7 +45,7 @@ class InMemColumnChunk {
     inline uint64_t getNumBytesPerValue() const { return numBytesPerValue; }
     inline uint64_t getNumBytes() const { return numBytes; }
     inline InMemColumnChunk* getNullChunk() { return nullChunk.get(); }
-    void copyChunk(std::shared_ptr<arrow::RecordBatch> batch);
+    void copyArrowBatch(std::shared_ptr<arrow::RecordBatch> batch);
     virtual void copyArrowArray(arrow::Array& arrowArray, arrow::Array* nodeOffsets = nullptr);
     virtual void flush(common::FileInfo* walFileInfo);
 

diff --git a/src/storage/copier/node_copier.cpp b/src/storage/copier/node_copier.cpp
@@ -188,8 +188,8 @@ void NPYNodeCopier::executeInternal(std::unique_ptr<CopyMorsel> morsel) {
     std::vector<std::unique_ptr<InMemColumnChunk>> columnChunks(1);
     columnChunks[0] =
         columns[columnToCopy]->getInMemColumnChunk(morsel->tupleIdx, endNodeOffset, &copyDesc);
-    auto chunk = reader->readBlock(morsel->blockIdx);
-    columnChunks[0]->copyChunk(chunk);
+    auto batch = reader->readBlock(morsel->blockIdx);
+    columnChunks[0]->copyArrowBatch(batch);
     for (auto i = 0u; i < morsel->numTuples; i++) {
         columnChunks[0]->setValueAtPos(reader->getPointerToRow(morsel->tupleIdx + i), i);
     }

diff --git a/src/storage/copier/npy_reader.cpp b/src/storage/copier/npy_reader.cpp
@@ -227,11 +227,9 @@ std::shared_ptr<arrow::RecordBatch> NpyReader::readBlock(common::block_idx_t blo
     auto buffer =
         std::make_shared<arrow::Buffer>(rowPointer, CopyConstants::NUM_ROWS_PER_BLOCK_FOR_NPY);
     auto arrowType = getArrowType();
-    int64_t length = CopyConstants::NUM_ROWS_PER_BLOCK_FOR_NPY < getNumRows() - rowNumber ?
-                         CopyConstants::NUM_ROWS_PER_BLOCK_FOR_NPY :
-                         getNumRows() - rowNumber;
+    int64_t length = std::min(CopyConstants::NUM_ROWS_PER_BLOCK_FOR_NPY, getNumRows() - rowNumber);
     auto arr = std::make_shared<arrow::PrimitiveArray>(arrowType, length, buffer);
-    auto field = std::make_shared<arrow::Field>("name", arrowType);
+    auto field = std::make_shared<arrow::Field>(defaultFieldName, arrowType);
     auto schema =
         std::make_shared<arrow::Schema>(std::vector<std::shared_ptr<arrow::Field>>{field});
     std::shared_ptr<arrow::RecordBatch> result;

diff --git a/src/storage/in_mem_storage_structure/in_mem_column_chunk.cpp b/src/storage/in_mem_storage_structure/in_mem_column_chunk.cpp
@@ -48,8 +48,9 @@ uint32_t InMemColumnChunk::getDataTypeSizeInColumn(common::LogicalType& dataType
     }
 }
 
-void InMemColumnChunk::copyChunk(std::shared_ptr<arrow::RecordBatch> batch) {
-    copyArrowArray(*batch->column(0), nullptr);
+void InMemColumnChunk::copyArrowBatch(std::shared_ptr<arrow::RecordBatch> batch) {
+    assert(batch->num_columns() == 1);
+    copyArrowArray(*batch->column(0), nullptr /* nodeOffsets */);
 }
 
 void InMemColumnChunk::copyArrowArray(arrow::Array& arrowArray, arrow::Array* nodeOffsets) {