thevar1able/ClickHouse
1
0
Fork 0
mirror of https://github.com/ClickHouse/ClickHouse.git synced 2024-09-20 00:30:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

Remove code duplication, use simdjson and rapidjson instead of Poco

Browse Source
This commit is contained in:
avogar 2021-12-27 20:30:40 +03:00
parent e3dbfe6bf6
commit 26abf7aa62
9 changed files with 181 additions and 207 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
src/CMakeLists.txt
View File

@ -526,6 +526,14 @@ if (USE_BZIP2)
target_include_directories (clickhouse_common_io SYSTEM BEFORE PRIVATE ${BZIP2_INCLUDE_DIR})
endif()
if(USE_SIMDJSON)
dbms_target_link_libraries(PRIVATE simdjson)
endif()
if(USE_RAPIDJSON)
dbms_target_include_directories(SYSTEM PRIVATE ${RAPIDJSON_INCLUDE_DIR})
endif()
dbms_target_link_libraries(PUBLIC consistent-hashing)
include ("${ClickHouse_SOURCE_DIR}/cmake/add_check.cmake")

8
src/Formats/EscapingRuleUtils.cpp
View File

@ -288,13 +288,7 @@ DataTypePtr determineDataTypeByEscapingRule(const String & field, const FormatSe
return parsed ? type : nullptr;
}
case FormatSettings::EscapingRule::JSON:
{
Poco::JSON::Parser parser;
String json = "{\"field\" : " + field + "}";
auto var = parser.parse(json);
Poco::JSON::Object::Ptr object = var.extract<Poco::JSON::Object::Ptr>();
return getDataTypeFromJSONField(object->get("field"));
}
return getDataTypeFromJSONField(field);
case FormatSettings::EscapingRule::CSV:
{
if (field.empty() || field == format_settings.csv.null_representation)

216
src/Formats/JSONEachRowUtils.cpp
View File

@ -9,10 +9,14 @@
#include <DataTypes/DataTypeArray.h>
#include <DataTypes/DataTypeTuple.h>
#include <DataTypes/DataTypeMap.h>
#include <Poco/JSON/Parser.h>
#include <Functions/SimdJSONParser.h>
#include <Functions/RapidJSONParser.h>
#include <Functions/DummyJSONParser.h>
#include <base/find_symbols.h>
#include <base/logger_useful.h>
namespace DB
{
namespace ErrorCodes
@ -34,7 +38,7 @@ static std::pair<bool, size_t> fileSegmentationEngineJSONEachRowImpl(ReadBuffer
while (loadAtPosition(in, memory, pos) && (balance || memory.size() + static_cast<size_t>(pos - in.position()) < min_chunk_size || number_of_rows < min_rows))
{
const auto current_object_size = memory.size() + static_cast<size_t>(pos - in.position());
if (current_object_size > 10 * min_chunk_size)
if (min_chunk_size != 0 && current_object_size > 10 * min_chunk_size)
throw ParsingException("Size of JSON object is extremely large. Expected not greater than " +
std::to_string(min_chunk_size) + " bytes, but current is " + std::to_string(current_object_size) +
" bytes per row. Increase the value setting 'min_chunk_bytes_for_parallel_parsing' or check your data manually, most likely JSON is malformed", ErrorCodes::INCORRECT_DATA);
@ -103,87 +107,20 @@ static std::pair<bool, size_t> fileSegmentationEngineJSONEachRowImpl(ReadBuffer
template <const char opening_bracket, const char closing_bracket>
static String readJSONEachRowLineIntoStringImpl(ReadBuffer & in)
{
skipWhitespaceIfAny(in);
if (in.eof())
throw Exception(ErrorCodes::INCORRECT_DATA, "Cannot read JSON object: unexpected end of file");
char * pos = in.position();
if (*pos != opening_bracket)
throw Exception(ErrorCodes::INCORRECT_DATA, "Cannot read JSONEachRow line: {} expected, {} got", opening_bracket, *in.position());
++pos;
Memory memory;
size_t balance = 1;
bool quotes = false;
while (loadAtPosition(in, memory, pos) && balance)
{
if (quotes)
{
pos = find_first_symbols<'\\', '"'>(pos, in.buffer().end());
if (pos == in.buffer().end())
continue;
if (*pos == '\\')
{
++pos;
if (loadAtPosition(in, memory, pos))
++pos;
}
else if (*pos == '"')
{
++pos;
quotes = false;
}
}
else
{
pos = find_first_symbols<opening_bracket, closing_bracket, '\\', '"'>(pos, in.buffer().end());
if (pos == in.buffer().end())
continue;
else if (*pos == opening_bracket)
{
++balance;
++pos;
}
else if (*pos == closing_bracket)
{
--balance;
++pos;
}
else if (*pos == '\\')
{
++pos;
if (loadAtPosition(in, memory, pos))
++pos;
}
else if (*pos == '"')
{
quotes = true;
++pos;
}
}
}
if (balance)
throw Exception(ErrorCodes::INCORRECT_DATA, "Cannot read JSON object: unexpected end of file");
saveUpToPosition(in, memory, pos);
fileSegmentationEngineJSONEachRowImpl<opening_bracket, closing_bracket>(in, memory, 0, 1);
return String(memory.data(), memory.size());
}
DataTypePtr getDataTypeFromJSONField(const Poco::Dynamic::Var & field)
DataTypePtr getDataTypeFromJSONFieldImpl(const SimdJSONParser::Element & field)
{
if (field.isEmpty())
if (field.isNull())
return nullptr;
if (field.isBoolean())
if (field.isBool())
return makeNullable(std::make_shared<DataTypeUInt8>());
if (field.isNumeric())
if (field.isInt64() || field.isUInt64() || field.isDouble())
return makeNullable(std::make_shared<DataTypeFloat64>());
if (field.isString())
@ -191,18 +128,18 @@ DataTypePtr getDataTypeFromJSONField(const Poco::Dynamic::Var & field)
if (field.isArray())
{
Poco::JSON::Array::Ptr array = field.extract<Poco::JSON::Array::Ptr>();
auto array = field.getArray();
/// Return nullptr in case of empty array because we cannot determine nested type.
if (array->size() == 0)
if (array.size() == 0)
return nullptr;
DataTypes nested_data_types;
/// If this array contains fields with different types we will treat it as Tuple.
bool is_tuple = false;
for (size_t i = 0; i != array->size(); ++i)
for (const auto element : array)
{
auto type = getDataTypeFromJSONField(array->get(i));
auto type = getDataTypeFromJSONFieldImpl(element);
if (!type)
return nullptr;
@ -218,14 +155,13 @@ DataTypePtr getDataTypeFromJSONField(const Poco::Dynamic::Var & field)
return std::make_shared<DataTypeArray>(nested_data_types.back());
}
if (field.type() == typeid(Poco::JSON::Object::Ptr))
if (field.isObject())
{
Poco::JSON::Object::Ptr object = field.extract<Poco::JSON::Object::Ptr>();
auto names = object->getNames();
auto object = field.getObject();
DataTypePtr value_type;
for (const auto & name : names)
for (const auto key_value_pair : object)
{
auto type = getDataTypeFromJSONField(object->get(name));
auto type = getDataTypeFromJSONFieldImpl(key_value_pair.second);
if (!type)
return nullptr;
@ -237,23 +173,45 @@ DataTypePtr getDataTypeFromJSONField(const Poco::Dynamic::Var & field)
return std::make_shared<DataTypeMap>(std::make_shared<DataTypeString>(), value_type);
}
throw Exception{ErrorCodes::INCORRECT_DATA, "Unexpected JSON type {}", field.type().name()};
throw Exception{ErrorCodes::INCORRECT_DATA, "Unexpected JSON type"};
}
using JSONEachRowFieldExtractor = std::function<std::vector<Poco::Dynamic::Var>(const Poco::Dynamic::Var &)>;
template <const char opening_bracket, const char closing_bracket>
static DataTypes determineColumnDataTypesFromJSONEachRowDataImpl(ReadBuffer & in, bool /*json_strings*/, JSONEachRowFieldExtractor extractor)
auto getJSONParserAndElement()
{
Poco::JSON::Parser parser;
DataTypes data_types;
#if USE_SIMDJSON
return std::pair<SimdJSONParser, SimdJSONParser::Element>();
#elif USE_RAPIDJSON
return std::pair<RapidJSONParser, RapidJSONParser::Element>();
#else
return std::pair<DummyJSONParser, DummyJSONParser::Element>();
#endif
}
DataTypePtr getDataTypeFromJSONField(const String & field)
{
auto [parser, element] = getJSONParserAndElement();
bool parsed = parser.parse(field, element);
if (!parsed)
throw Exception(ErrorCodes::INCORRECT_DATA, "Cannot parse JSON object");
return getDataTypeFromJSONFieldImpl(element);
}
template <class Extractor, const char opening_bracket, const char closing_bracket>
static DataTypes determineColumnDataTypesFromJSONEachRowDataImpl(ReadBuffer & in, bool /*json_strings*/, Extractor & extractor)
{
String line = readJSONEachRowLineIntoStringImpl<opening_bracket, closing_bracket>(in);
auto var = parser.parse(line);
std::vector<Poco::Dynamic::Var> fields = extractor(var);
auto [parser, element] = getJSONParserAndElement();
bool parsed = parser.parse(line, element);
if (!parsed)
throw Exception(ErrorCodes::INCORRECT_DATA, "Cannot parse JSON object");
auto fields = extractor.extract(element);
DataTypes data_types;
data_types.reserve(fields.size());
for (const auto & field : fields)
data_types.push_back(getDataTypeFromJSONField(field));
data_types.push_back(getDataTypeFromJSONFieldImpl(field));
/// TODO: For JSONStringsEachRow/JSONCompactStringsEach all types will be strings.
/// Should we try to parse data inside strings somehow in this case?
@ -271,43 +229,57 @@ std::pair<bool, size_t> fileSegmentationEngineJSONCompactEachRow(ReadBuffer & in
return fileSegmentationEngineJSONEachRowImpl<'[', ']'>(in, memory, min_chunk_size, min_rows);
}
struct JSONEachRowFieldsExtractor
{
template <class Element>
std::vector<Element> extract(const Element & element)
{
/// {..., "<column_name>" : <value>, ...}
auto object = element.getObject();
std::vector<Element> fields;
fields.reserve(object.size());
column_names.reserve(object.size());
for (const auto & key_value_pair : object)
{
column_names.emplace_back(key_value_pair.first);
fields.push_back(key_value_pair.second);
}
return fields;
}
std::vector<String> column_names;
};
std::unordered_map<String, DataTypePtr> readRowAndGetNamesAndDataTypesForJSONEachRow(ReadBuffer & in, bool json_strings)
{
std::vector<String> column_names;
/// {..., "<column_name>" : <value>, ...}
auto extractor = [&](const Poco::Dynamic::Var & var)
{
Poco::JSON::Object::Ptr object = var.extract<Poco::JSON::Object::Ptr>();
column_names = object->getNames();
std::vector<Poco::Dynamic::Var> fields;
for (size_t i = 0; i != object->size(); ++i)
fields.push_back(object->get(column_names[i]));
return fields;
};
auto data_types = determineColumnDataTypesFromJSONEachRowDataImpl<'{', '}'>(in, json_strings, extractor);
JSONEachRowFieldsExtractor extractor;
auto data_types = determineColumnDataTypesFromJSONEachRowDataImpl<JSONEachRowFieldsExtractor, '{', '}'>(in, json_strings, extractor);
std::unordered_map<String, DataTypePtr> result;
for (size_t i = 0; i != column_names.size(); ++i)
result[column_names[i]] = data_types[i];
for (size_t i = 0; i != extractor.column_names.size(); ++i)
result[extractor.column_names[i]] = data_types[i];
return result;
}
struct JSONCompactEachRowFieldsExtractor
{
template <class Element>
std::vector<Element> extract(const Element & element)
{
/// [..., <value>, ...]
auto array = element.getArray();
std::vector<Element> fields;
fields.reserve(array.size());
for (size_t i = 0; i != array.size(); ++i)
fields.push_back(array[i]);
return fields;
}
};
DataTypes readRowAndGetDataTypesForJSONCompactEachRow(ReadBuffer & in, bool json_strings)
{
/// [..., <value>, ...]
auto extractor = [](const Poco::Dynamic::Var & var)
{
Poco::JSON::Array::Ptr array = var.extract<Poco::JSON::Array::Ptr>();
std::vector<Poco::Dynamic::Var> fields;
fields.reserve(array->size());
for (size_t i = 0; i != array->size(); ++i)
fields.push_back(array->get(i));
return fields;
};
return determineColumnDataTypesFromJSONEachRowDataImpl<'[', ']'>(in, json_strings, extractor);
JSONCompactEachRowFieldsExtractor extractor;
return determineColumnDataTypesFromJSONEachRowDataImpl<JSONCompactEachRowFieldsExtractor, '[', ']'>(in, json_strings, extractor);
}

7
src/Formats/JSONEachRowUtils.h
View File

@ -1,9 +1,10 @@
#pragma once
#include "config_formats.h"
#include <Formats/FormatSettings.h>
#include <IO/BufferWithOwnMemory.h>
#include <IO/ReadBuffer.h>
#include <Poco/JSON/Object.h>
#include <utility>
namespace DB
@ -13,10 +14,10 @@ std::pair<bool, size_t> fileSegmentationEngineJSONEachRow(ReadBuffer & in, DB::M
std::pair<bool, size_t> fileSegmentationEngineJSONCompactEachRow(ReadBuffer & in, DB::Memory<> & memory, size_t min_chunk_size, size_t min_rows);
/// Convert JSON type to ClickHouse type. Make the result type always Nullable.
/// Parse JSON from string and convert it's type to ClickHouse type. Make the result type always Nullable.
/// JSON array with different nested types is treated as Tuple.
/// If cannot convert (for example when field contains null), return nullptr.
DataTypePtr getDataTypeFromJSONField(const Poco::Dynamic::Var & field);
DataTypePtr getDataTypeFromJSONField(const String & field);
/// Read row in JSONEachRow format and try to determine type for each field.
/// Return map {column_name : type}.

3
src/Formats/config_formats.h.in
View File

@ -10,4 +10,5 @@
#cmakedefine01 USE_ARROW
#cmakedefine01 USE_PROTOBUF
#cmakedefine01 USE_MSGPACK
#cmakedefine01 USE_SIMDJSON
#cmakedefine01 USE_RAPIDJSON

4
src/Processors/Formats/ISchemaReader.cpp
View File

@ -55,12 +55,12 @@ NamesAndTypesList IRowSchemaReader::readSchema()
if (data_types.empty())
throw Exception(ErrorCodes::CANNOT_EXTRACT_TABLE_STRUCTURE, "Cannot read rows from the data");
/// If column names weren't set, use default names 'column_1', 'column_2', ...
/// If column names weren't set, use default names 'c1', 'c2', ...
if (column_names.empty())
{
column_names.reserve(data_types.size());
for (size_t i = 0; i != data_types.size(); ++i)
column_names.push_back("column_" + std::to_string(i + 1));
column_names.push_back("c" + std::to_string(i + 1));
}
/// If column names were set, check that the number of names match the number of types.
else if (column_names.size() != data_types.size())

15
src/Processors/Formats/Impl/ORCBlockInputFormat.cpp
View File

@ -88,9 +88,13 @@ static size_t countIndicesForType(std::shared_ptr<arrow::DataType> type)
}
static void getFileReaderAndSchema(
ReadBuffer & in, std::unique_ptr<arrow::adapters::orc::ORCFileReader> & file_reader, std::shared_ptr<arrow::Schema> & schema, const FormatSettings & format_settings)
ReadBuffer & in,
std::unique_ptr<arrow::adapters::orc::ORCFileReader> & file_reader,
std::shared_ptr<arrow::Schema> & schema,
const FormatSettings & format_settings,
std::atomic<int> & is_stopped)
{
auto arrow_file = asArrowFile(*in, format_settings, is_stopped);
auto arrow_file = asArrowFile(in, format_settings, is_stopped);
if (is_stopped)
return;
@ -108,7 +112,9 @@ static void getFileReaderAndSchema(
void ORCBlockInputFormat::prepareReader()
{
std::shared_ptr<arrow::Schema> schema;
getFileReaderAndSchema(*in, file_reader, schema, format_settings);
getFileReaderAndSchema(*in, file_reader, schema, format_settings, is_stopped);
if (is_stopped)
return;
arrow_column_to_ch_column = std::make_unique<ArrowColumnToCHColumn>(getPort().getHeader(), "ORC", format_settings.orc.import_nested);
@ -143,7 +149,8 @@ NamesAndTypesList ORCSchemaReader::readSchema()
{
std::unique_ptr<arrow::adapters::orc::ORCFileReader> file_reader;
std::shared_ptr<arrow::Schema> schema;
getFileReaderAndSchema(in, file_reader, schema, format_settings);
std::atomic<int> is_stopped = 0;
getFileReaderAndSchema(in, file_reader, schema, format_settings, is_stopped);
auto header = ArrowColumnToCHColumn::arrowSchemaToCHHeader(*schema, "ORC");
return header.getNamesAndTypesList();
}

13
src/Storages/HDFS/StorageHDFS.cpp
View File

@ -117,9 +117,6 @@ namespace
}
}
static Strings listFilesWithRegexpMatching(const String & path_for_ls, const HDFSFSPtr & fs, const String & for_match);
StorageHDFS::StorageHDFS(
const String & uri_,
const StorageID & table_id_,
@ -187,14 +184,8 @@ class HDFSSource::DisclosedGlobIterator::Impl
public:
Impl(ContextPtr context_, const String & uri)
{
const size_t begin_of_path = uri.find('/', uri.find("//") + 2);
const String path_from_uri = uri.substr(begin_of_path);
const String uri_without_path = uri.substr(0, begin_of_path); /// ends without '/'
HDFSBuilderWrapper builder = createHDFSBuilder(uri_without_path + "/", context_->getGlobalContext()->getConfigRef());
HDFSFSPtr fs = createHDFSFS(builder.get());
uris = listFilesWithRegexpMatching("/", fs, path_from_uri);
const auto [path_from_uri, uri_without_path] = getPathFromUriAndUriWithoutPath(uri);
uris = getPathsList(path_from_uri, uri_without_path, context_);
for (auto & elem : uris)
elem = uri_without_path + elem;
uris_iter = uris.begin();

114
tests/queries/0_stateless/02149_schema_inference.reference
View File

@ -1,8 +1,8 @@
TSV
column_1 Nullable(String)
column_2 Nullable(String)
column_3 Nullable(String)
column_4 Nullable(String)
c1 Nullable(String)
c2 Nullable(String)
c3 Nullable(String)
c4 Nullable(String)
42 Some string [1, 2, 3, 4] (1, 2, 3)
42 abcd [] (4, 5, 6)
TSVWithNames
@ -13,18 +13,18 @@ tuple Nullable(String)
42 Some string [1, 2, 3, 4] (1, 2, 3)
42 abcd [] (4, 5, 6)
CSV
column_1 Nullable(Float64)
column_2 Nullable(String)
column_3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
column_4 Array(Nullable(Float64))
c1 Nullable(Float64)
c2 Nullable(String)
c3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
c4 Array(Nullable(Float64))
\N Some string [([1,2.3],'String'),([],NULL)] [1,NULL,3]
42 \N [([1,2.3],'String'),([3],'abcd')] [4,5,6]
column_1 Nullable(String)
column_2 Nullable(String)
c1 Nullable(String)
c2 Nullable(String)
42 String
String 42
column_1 Nullable(String)
column_2 Nullable(String)
c1 Nullable(String)
c2 Nullable(String)
\N [NULL, NULL]
\N []
CSVWithNames
@ -35,15 +35,15 @@ d Array(Nullable(Float64))
\N Some string [([1,2.3],'String'),([],NULL)] [1,NULL,3]
42 \N [([1,2.3],'String'),([3],'abcd')] [4,5,6]
JSONCompactEachRow
column_1 Nullable(Float64)
column_2 Array(Tuple(Nullable(Float64), Nullable(String)))
column_3 Map(String, Nullable(Float64))
column_4 Nullable(UInt8)
c1 Nullable(Float64)
c2 Array(Tuple(Nullable(Float64), Nullable(String)))
c3 Map(String, Nullable(Float64))
c4 Nullable(UInt8)
42.42 [(1,'String'),(2,'abcd')] {'key':42,'key2':24} 1
column_1 Nullable(Float64)
column_2 Array(Tuple(Nullable(Float64), Nullable(String)))
column_3 Map(String, Nullable(Float64))
column_4 Nullable(UInt8)
c1 Nullable(Float64)
c2 Array(Tuple(Nullable(Float64), Nullable(String)))
c3 Map(String, Nullable(Float64))
c4 Nullable(UInt8)
\N [(1,'String'),(2,NULL)] {'key':NULL,'key2':24} \N
32 [(2,'String 2'),(3,'hello')] {'key3':4242,'key4':2424} 1
JSONCompactEachRowWithNames
@ -82,60 +82,60 @@ s1 \N 1
\N \N \N
\N [3] \N
Values
column_1 Nullable(Float64)
column_2 Nullable(String)
column_3 Array(Nullable(Float64))
column_4 Tuple(Nullable(Float64), Nullable(String))
column_5 Tuple(Array(Nullable(Float64)), Array(Tuple(Nullable(Float64), Nullable(String))))
c1 Nullable(Float64)
c2 Nullable(String)
c3 Array(Nullable(Float64))
c4 Tuple(Nullable(Float64), Nullable(String))
c5 Tuple(Array(Nullable(Float64)), Array(Tuple(Nullable(Float64), Nullable(String))))
42.42 Some string [1,2,3] (1,'2') ([1,2],[(3,'4'),(5,'6')])
column_1 Nullable(Float64)
column_2 Nullable(String)
column_3 Array(Nullable(Float64))
column_4 Tuple(Nullable(Float64), Nullable(Float64))
column_5 Tuple(Array(Nullable(Float64)), Array(Tuple(Nullable(Float64), Nullable(String))))
c1 Nullable(Float64)
c2 Nullable(String)
c3 Array(Nullable(Float64))
c4 Tuple(Nullable(Float64), Nullable(Float64))
c5 Tuple(Array(Nullable(Float64)), Array(Tuple(Nullable(Float64), Nullable(String))))
42.42 \N [1,NULL,3] (1,NULL) ([1,2],[(3,'4'),(5,'6')])
\N Some string [10] (1,2) ([],[])
Regexp
column_1 Nullable(String)
column_2 Nullable(String)
column_3 Nullable(String)
c1 Nullable(String)
c2 Nullable(String)
c3 Nullable(String)
42 Some string 1 [([1, 2, 3], String 1), ([], String 1)]
2 Some string 2 [([4, 5, 6], String 2), ([], String 2)]
312 Some string 3 [([1, 2, 3], String 2), ([], String 2)]
column_1 Nullable(Float64)
column_2 Nullable(String)
column_3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
c1 Nullable(Float64)
c2 Nullable(String)
c3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
42 Some string 1 [([1,2,3],'String 1'),([],'String 1')]
3 Some string 2 [([3,5,1],'String 2'),([],'String 2')]
244 Some string 3 [([],'String 3'),([],'String 3')]
column_1 Nullable(Float64)
column_2 Nullable(String)
column_3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
c1 Nullable(Float64)
c2 Nullable(String)
c3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
42 Some string 1 [([1,2,3],'String 1'),([],'String 1')]
2 Some string 2 [([],'String 2'),([],'String 2')]
43 Some string 3 [([1,5,3],'String 3'),([],'String 3')]
column_1 Nullable(Float64)
column_2 Nullable(String)
column_3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
c1 Nullable(Float64)
c2 Nullable(String)
c3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
42 Some string 1 [([1,2,3],'String 1'),([1],'String 1')]
52 Some string 2 [([],'String 2'),([1],'String 2')]
24 Some string 3 [([1,2,3],'String 3'),([1],'String 3')]
CustomSeparated
column_1 Nullable(Float64)
column_2 Nullable(String)
column_3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
c1 Nullable(Float64)
c2 Nullable(String)
c3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
42.42 Some string 1 [([1,2,3],'String 1'),([1],'String 1')]
42 Some string 2 [([],'String 2'),([],'String 2')]
\N Some string 3 [([1,2,3],'String 3'),([1],'String 3')]
column_1 Nullable(Float64)
column_2 Nullable(String)
column_3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
c1 Nullable(Float64)
c2 Nullable(String)
c3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
42.42 Some string 1 [([1,2,3],'String 1'),([1],'String 1')]
42 Some string 2 [([],'String 2'),([],'String 2')]
\N Some string 3 [([1,2,3],'String 3'),([1],'String 3')]
column_1 Nullable(Float64)
column_2 Nullable(String)
column_3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
c1 Nullable(Float64)
c2 Nullable(String)
c3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
42.42 Some string 1 [([1,2,3],'String 1'),([1],'String 1')]
42 Some string 2 [([],'String 2'),([],'String 2')]
\N Some string 3 [([1,2,3],'String 3'),([1],'String 3')]
@ -159,12 +159,12 @@ column_3 Array(Tuple(Array(Nullable(Float64)), Nullable(String)))
42 Some string 2 [([],'String 2'),([],'String 2')]
\N Some string 3 [([1,2,3],'String 3'),([1],'String 3')]
MsgPack
column_1 Nullable(Int64)
column_2 Nullable(Int64)
column_3 Nullable(Float32)
column_4 Nullable(String)
column_5 Array(Array(Nullable(Int64)))
column_6 Map(Int64, Array(Nullable(Int64)))
c1 Nullable(Int64)
c2 Nullable(Int64)
c3 Nullable(Float32)
c4 Nullable(String)
c5 Array(Array(Nullable(Int64)))
c6 Map(Int64, Array(Nullable(Int64)))
\N 0 0 Str: 0 [[0,1],[0]] {0:[0,1]}
1 \N 1 Str: 1 [[1,2],[1]] {1:[1,2]}
\N 2 2 Str: 2 [[2,3],[2]] {2:[2,3]}