| // 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 PARQUET_DELTA_BIT_PACK_ENCODING_H |
| #define PARQUET_DELTA_BIT_PACK_ENCODING_H |
| |
| #include "encodings.h" |
| |
| namespace parquet_cpp { |
| |
| template<int num_values, int bit_width> |
| void DecodeMiniBlock(uint8_t* data) { |
| } |
| |
| class DeltaBitPackDecoder : public Decoder { |
| public: |
| DeltaBitPackDecoder(const parquet::Type::type& type) |
| : Decoder(type, parquet::Encoding::DELTA_BINARY_PACKED) { |
| if (type != parquet::Type::INT32 && type != parquet::Type::INT64) { |
| throw ParquetException("Delta bit pack encoding should only be for integer data."); |
| } |
| } |
| |
| virtual void SetData(int num_values, const uint8_t* data, int len) { |
| num_values_ = num_values; |
| decoder_ = impala::BitReader(data, len); |
| values_current_block_ = 0; |
| values_current_mini_block_ = 0; |
| } |
| |
| virtual int GetInt32(int32_t* buffer, int max_values) { |
| return GetInternal(buffer, max_values); |
| } |
| |
| virtual int GetInt64(int64_t* buffer, int max_values) { |
| return GetInternal(buffer, max_values); |
| } |
| |
| static void Init() { |
| Init(decode_data_14_, 14); |
| Init(decode_data_15_, 15); |
| } |
| |
| private: |
| void InitBlock() { |
| uint64_t block_size; |
| if (!decoder_.GetVlqInt(&block_size)) ParquetException::EofException(); |
| if (!decoder_.GetVlqInt(&num_mini_blocks_)) ParquetException::EofException(); |
| if (!decoder_.GetVlqInt(&values_current_block_)) { |
| ParquetException::EofException(); |
| } |
| if (!decoder_.GetZigZagVlqInt(&last_value_)) ParquetException::EofException(); |
| delta_bit_widths_.resize(num_mini_blocks_); |
| |
| if (!decoder_.GetZigZagVlqInt(&min_delta_)) ParquetException::EofException(); |
| for (int i = 0; i < num_mini_blocks_; ++i) { |
| if (!decoder_.GetAligned<uint8_t>(1, &delta_bit_widths_[i])) { |
| ParquetException::EofException(); |
| } |
| } |
| values_per_mini_block_ = block_size / num_mini_blocks_; |
| mini_block_idx_ = 0; |
| delta_bit_width_ = delta_bit_widths_[0]; |
| values_current_mini_block_ = values_per_mini_block_; |
| } |
| |
| template <typename T> |
| int GetInternal(T* buffer, int max_values) { |
| max_values = std::min(max_values, num_values_); |
| for (int i = 0; i < max_values; ++i) { |
| if (UNLIKELY(values_current_mini_block_ == 0)) { |
| ++mini_block_idx_; |
| if (mini_block_idx_ < delta_bit_widths_.size()) { |
| delta_bit_width_ = delta_bit_widths_[mini_block_idx_]; |
| values_current_mini_block_ = values_per_mini_block_; |
| } else { |
| InitBlock(); |
| buffer[i] = last_value_; |
| continue; |
| } |
| } |
| |
| if (max_values == 32) { |
| if (delta_bit_width_ == 14) { |
| const uint8_t* data = decoder_.current_ptr(); |
| DecodeBlock<T>(decode_data_14_, data, buffer); |
| } else if (delta_bit_width_ == 15) { |
| const uint8_t* data = decoder_.current_ptr(); |
| DecodeBlock<T>(decode_data_15_, data, buffer); |
| } else { |
| // TODO: the key to this algorithm is to decode the entire miniblock at once. |
| ParquetException::NYI("miniblocks"); |
| } |
| for (int i = 0; i < 32; ++i) { |
| buffer[i] = last_value_ + min_delta_ + buffer[i]; |
| last_value_ = buffer[i]; |
| } |
| decoder_.SkipBytes(32 * delta_bit_width_ / 8); |
| values_current_mini_block_ = 0; |
| num_values_ -= 32; |
| return 32; |
| } |
| int64_t delta; |
| if (!decoder_.GetValue(delta_bit_width_, &delta)) ParquetException::EofException(); |
| delta += min_delta_; |
| last_value_ += delta; |
| buffer[i] = last_value_; |
| --values_current_mini_block_; |
| } |
| num_values_ -= max_values; |
| return max_values; |
| } |
| |
| impala::BitReader decoder_; |
| uint64_t values_current_block_; |
| uint64_t num_mini_blocks_; |
| int values_per_mini_block_; |
| int values_current_mini_block_; |
| |
| int64_t min_delta_; |
| int mini_block_idx_; |
| std::vector<uint8_t> delta_bit_widths_; |
| int delta_bit_width_; |
| |
| int64_t last_value_; |
| |
| struct DecodeData { |
| int byte_offset; |
| int shift; |
| }; |
| |
| static DecodeData decode_data_14_[32]; |
| static DecodeData decode_data_15_[32]; |
| |
| template <typename T> |
| void DecodeBlock(const DecodeData* offsets, const uint8_t* data, T* buffer) { |
| uint64_t mask = -1; |
| mask = mask >> (64 - delta_bit_width_); |
| for (int i = 0; i < 32; i += 4) { |
| buffer[i + 0] = *reinterpret_cast<const T*>(data + offsets[i + 0].byte_offset); |
| buffer[i + 1] = *reinterpret_cast<const T*>(data + offsets[i + 1].byte_offset); |
| buffer[i + 2] = *reinterpret_cast<const T*>(data + offsets[i + 2].byte_offset); |
| buffer[i + 3] = *reinterpret_cast<const T*>(data + offsets[i + 3].byte_offset); |
| |
| buffer[i + 0] >>= offsets[i + 0].shift; |
| buffer[i + 1] >>= offsets[i + 1].shift; |
| buffer[i + 2] >>= offsets[i + 2].shift; |
| buffer[i + 3] >>= offsets[i + 3].shift; |
| |
| buffer[i + 0] &= mask; |
| buffer[i + 1] &= mask; |
| buffer[i + 2] &= mask; |
| buffer[i + 3] &= mask; |
| } |
| } |
| |
| static void Init(DecodeData* data, int bit_width) { |
| int bit_offset = 0; |
| for (int i = 0; i < 32; ++i) { |
| data[i].byte_offset = bit_offset / 8; |
| data[i].shift = bit_offset % 8; |
| bit_offset += bit_width; |
| } |
| } |
| }; |
| |
| } |
| |
| #endif |
| |