#pragma once #include "config_formats.h" #if USE_ORC #include #include #include #include namespace arrow::adapters::orc { class ORCFileReader; } namespace DB { class ArrowColumnToCHColumn; class ORCBlockInputFormat : public IInputFormat { public: ORCBlockInputFormat(ReadBuffer & in_, Block header_, const FormatSettings & format_settings_); String getName() const override { return "ORCBlockInputFormat"; } void resetParser() override; const BlockMissingValues & getMissingValues() const override; protected: Chunk generate() override; void onCancel() override { is_stopped = 1; } private: void prepareReader(); // TODO: check that this class implements every part of its parent std::unique_ptr file_reader; std::unique_ptr arrow_column_to_ch_column; // indices of columns to read from ORC file std::vector include_indices; std::vector missing_columns; BlockMissingValues block_missing_values; const FormatSettings format_settings; const std::unordered_set & skip_stripes; int stripe_total = 0; int stripe_current = 0; std::atomic is_stopped{0}; }; class ORCSchemaReader : public ISchemaReader { public: ORCSchemaReader(ReadBuffer & in_, const FormatSettings & format_settings_); NamesAndTypesList readSchema() override; private: const FormatSettings format_settings; }; } #endif