ClickHouse/dbms/Storages/MergeTree/MergeTreeDataPartWriterWide.h

80 lines
2.5 KiB
C++
Raw Normal View History

2019-10-21 00:28:29 +00:00
#include <Storages/MergeTree/IMergeTreeDataPartWriter.h>
namespace DB
{
2020-02-03 12:08:40 +00:00
/// Writes data part in wide format.
2019-10-21 17:23:06 +00:00
class MergeTreeDataPartWriterWide : public IMergeTreeDataPartWriter
2019-10-21 00:28:29 +00:00
{
public:
2020-01-17 12:24:27 +00:00
2019-10-21 15:33:59 +00:00
using ColumnToSize = std::map<std::string, UInt64>;
MergeTreeDataPartWriterWide(
DiskPtr disk,
2019-10-21 15:33:59 +00:00
const String & part_path,
const MergeTreeData & storage,
const NamesAndTypesList & columns_list,
const std::vector<MergeTreeIndexPtr> & indices_to_recalc,
2019-10-21 15:33:59 +00:00
const String & marks_file_extension,
const CompressionCodecPtr & default_codec,
const MergeTreeWriterSettings & settings,
2019-11-07 11:11:38 +00:00
const MergeTreeIndexGranularity & index_granularity);
2019-10-21 15:33:59 +00:00
2020-03-18 03:27:32 +00:00
void write(const Block & block, const IColumn::Permutation * permutation,
const Block & primary_key_block, const Block & skip_indexes_block) override;
2019-10-21 00:28:29 +00:00
2020-03-18 03:27:32 +00:00
void finishDataSerialization(IMergeTreeDataPart::Checksums & checksums, bool sync) override;
2019-10-21 17:23:06 +00:00
2019-12-09 21:21:17 +00:00
IDataType::OutputStreamGetter createStreamGetter(const String & name, WrittenOffsetColumns & offset_columns);
2019-10-21 00:28:29 +00:00
2020-01-16 16:15:01 +00:00
private:
2019-10-21 00:28:29 +00:00
/// Write data of one column.
/// Return how many marks were written and
/// how many rows were written for last mark
void writeColumn(
2019-10-21 00:28:29 +00:00
const String & name,
const IDataType & type,
const IColumn & column,
2019-12-09 21:21:17 +00:00
WrittenOffsetColumns & offset_columns);
2019-10-21 00:28:29 +00:00
/// Write single granule of one column (rows between 2 marks)
size_t writeSingleGranule(
const String & name,
const IDataType & type,
const IColumn & column,
WrittenOffsetColumns & offset_columns,
IDataType::SerializeBinaryBulkStatePtr & serialization_state,
IDataType::SerializeBinaryBulkSettings & serialize_settings,
size_t from_row,
size_t number_of_rows,
bool write_marks);
/// Write mark for column
void writeSingleMark(
const String & name,
const IDataType & type,
WrittenOffsetColumns & offset_columns,
size_t number_of_rows,
DB::IDataType::SubstreamPath & path);
2019-10-21 17:23:06 +00:00
void writeFinalMark(
const std::string & column_name,
const DataTypePtr column_type,
WrittenOffsetColumns & offset_columns,
DB::IDataType::SubstreamPath & path);
2019-10-21 00:28:29 +00:00
2019-10-21 15:33:59 +00:00
void addStreams(
const String & name,
const IDataType & type,
const CompressionCodecPtr & effective_codec,
2019-12-09 21:21:17 +00:00
size_t estimated_size);
2019-10-21 15:33:59 +00:00
2019-10-21 00:28:29 +00:00
SerializationStates serialization_states;
2020-02-19 14:07:36 +00:00
using ColumnStreams = std::map<String, StreamPtr>;
2019-10-21 00:28:29 +00:00
ColumnStreams column_streams;
};
2019-10-21 15:33:59 +00:00
}