hdfs-parquet-table-writer.h

Go to the documentation of this file.

// Copyright 2012 Cloudera Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.


#ifndef IMPALA_EXEC_HDFS_PARQUET_TABLE_WRITER_H
#define IMPALA_EXEC_HDFS_PARQUET_TABLE_WRITER_H

#include "exec/data-sink.h"

#include <hdfs.h>
#include <map>
#include <boost/scoped_ptr.hpp>

#include "util/compress.h"
#include "runtime/descriptors.h"
#include "exec/hdfs-table-writer.h"
#include "exec/parquet-common.h"

namespace impala {

class Expr;
struct OutputPartition;
class RuntimeState;
class ThriftSerializer;
class TupleRow;


class HdfsParquetTableWriter : public HdfsTableWriter {
 public:
  HdfsParquetTableWriter(HdfsTableSink* parent,
                         RuntimeState* state, OutputPartition* output_partition,
                         const HdfsPartitionDescriptor* part_desc,
                         const HdfsTableDescriptor* table_desc,
                         const std::vector<ExprContext*>& output_expr_ctxs);

  ~HdfsParquetTableWriter();

  virtual Status Init();

  virtual Status InitNewFile();

  virtual Status AppendRowBatch(RowBatch* batch,
                                const std::vector<int32_t>& row_group_indices,
                                bool* new_file);

  virtual Status Finalize();

  virtual void Close();

  virtual uint64_t default_block_size() const;

  virtual std::string file_extension() const { return "parq"; }

 private:
  static const int DEFAULT_DATA_PAGE_SIZE = 64 * 1024;

  static const int64_t MAX_DATA_PAGE_SIZE = 1024 * 1024 * 1024;

  static const int HDFS_BLOCK_SIZE = 256 * 1024 * 1024;

  static const int HDFS_BLOCK_ALIGNMENT = 1024 * 1024;

  static const int ROW_GROUP_SIZE = HDFS_BLOCK_SIZE;

  static const int HDFS_MIN_FILE_SIZE = 8 * 1024 * 1024;

  class BaseColumnWriter;
  friend class BaseColumnWriter;

  template<typename T> class ColumnWriter;
  template<typename T> friend class ColumnWriter;
  class BoolColumnWriter;
  friend class BoolColumnWriter;

  int64_t MinBlockSize() const;

  Status CreateSchema();

  Status WriteFileHeader();

  Status WriteFileFooter();

  Status FlushCurrentRowGroup();

  Status AddRowGroup();

  boost::scoped_ptr<ThriftSerializer> thrift_serializer_;

  parquet::FileMetaData file_metadata_;

  parquet::RowGroup* current_row_group_;

  std::vector<BaseColumnWriter*> columns_;

  int64_t row_count_;

  int64_t file_size_estimate_;

  int64_t file_size_limit_;

  int64_t file_pos_;

  boost::scoped_ptr<MemPool> reusable_col_mem_pool_;

  boost::scoped_ptr<MemPool> per_file_mem_pool_;

  int row_idx_;

  std::vector<uint8_t> compression_staging_buffer_;

  TParquetInsertStats parquet_stats_;
};

}
#endif