dbms: Server: Merged with master. [#METR-18844]

contrib/libpoco/CMakeLists.txt

@@ -136,8 +136,8 @@ endif(WIN32)
 
 if (UNIX AND NOT ANDROID )
   add_definitions( -DPOCO_OS_FAMILY_UNIX )
-  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-unused-private-field -Wno-unused-local-typedef -Wno-for-loop-analysis -Wno-unknown-pragmas")
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-unused-private-field -Wno-unused-local-typedef -Wno-for-loop-analysis -Wno-unknown-pragmas -Wno-unused-variable")
-  set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Wno-unknown-pragmas")
+  set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Wno-unknown-pragmas -Wno-unused-variable")
   # Standard 'must be' defines
   if (APPLE)
     add_definitions( -DPOCO_HAVE_IPv6 -DPOCO_NO_STAT64)

dbms/CMakeLists.txt

@@ -266,6 +266,9 @@ add_library (dbms
 	include/DB/DataStreams/CSVRowOutputStream.h
 	include/DB/DataStreams/CSVRowInputStream.h
 	include/DB/DataStreams/verbosePrintString.h
+	include/DB/DataStreams/SquashingTransform.h
+	include/DB/DataStreams/SquashingBlockInputStream.h
+	include/DB/DataStreams/SquashingBlockOutputStream.h
 	include/DB/DataTypes/IDataType.h
 	include/DB/DataTypes/IDataTypeDummy.h
 	include/DB/DataTypes/DataTypeSet.h
@@ -744,6 +747,9 @@ add_library (dbms
 	src/DataStreams/RemoteBlockInputStream.cpp
 	src/DataStreams/BlockIO.cpp
 	src/DataStreams/verbosePrintString.cpp
+	src/DataStreams/SquashingTransform.cpp
+	src/DataStreams/SquashingBlockInputStream.cpp
+	src/DataStreams/SquashingBlockOutputStream.cpp
 
 	src/DataTypes/DataTypeString.cpp
 	src/DataTypes/DataTypeFixedString.cpp

dbms/include/DB/Common/getNumberOfPhysicalCPUCores.h

@@ -1,4 +1,4 @@
 #pragma once
 
-/// Получить количество ядер CPU без учёта hyper-threading.
+/// Get number of CPU cores without hyper-threading.
 unsigned getNumberOfPhysicalCPUCores();

dbms/include/DB/DataStreams/SquashingBlockInputStream.h

@@ -0,0 +1,34 @@
+#pragma once
+
+#include <DB/DataStreams/IProfilingBlockInputStream.h>
+#include <DB/DataStreams/SquashingTransform.h>
+
+
+namespace DB
+{
+
+/** Merging consecutive blocks of stream to specified minimum size.
+  */
+class SquashingBlockInputStream : public IProfilingBlockInputStream
+{
+public:
+	SquashingBlockInputStream(BlockInputStreamPtr & src, size_t min_block_size_rows, size_t min_block_size_bytes);
+
+	String getName() const override { return "Squashing"; }
+
+	String getID() const override
+	{
+		std::stringstream res;
+		res << "Squashing(" << children.at(0)->getID() << ")";
+		return res.str();
+	}
+
+protected:
+	Block readImpl() override;
+
+private:
+	SquashingTransform transform;
+	bool all_read = false;
+};
+
+}

dbms/include/DB/DataStreams/SquashingBlockOutputStream.h

@@ -0,0 +1,32 @@
+#pragma once
+
+#include <DB/DataStreams/IBlockOutputStream.h>
+#include <DB/DataStreams/SquashingTransform.h>
+
+
+namespace DB
+{
+
+/** Merging consecutive blocks of stream to specified minimum size.
+  */
+class SquashingBlockOutputStream : public IBlockOutputStream
+{
+public:
+	SquashingBlockOutputStream(BlockOutputStreamPtr & dst, size_t min_block_size_rows, size_t min_block_size_bytes);
+
+	void write(const Block & block) override;
+
+	void flush() override;
+	void writePrefix() override;
+	void writeSuffix() override;
+
+private:
+	BlockOutputStreamPtr output;
+
+	SquashingTransform transform;
+	bool all_written = false;
+
+	void finalize();
+};
+
+}

dbms/include/DB/DataStreams/SquashingTransform.h

@@ -0,0 +1,52 @@
+#include <DB/Core/Block.h>
+
+
+namespace DB
+{
+
+
+/** Merging consecutive passed blocks to specified minimum size.
+  *
+  * (But if one of input blocks has already at least specified size,
+  *  then don't merge it with neighbours, even if neighbours are small.)
+  *
+  * Used to prepare blocks to adequate size for INSERT queries,
+  *  because such storages as Memory, StripeLog, Log, TinyLog...
+  *  store or compress data in blocks exactly as passed to it,
+  *  and blocks of small size are not efficient.
+  *
+  * Order of data is kept.
+  */
+class SquashingTransform
+{
+public:
+	/// Conditions on rows and bytes are OR-ed. If one of them is zero, then corresponding condition is ignored.
+	SquashingTransform(size_t min_block_size_rows, size_t min_block_size_bytes);
+
+	/// When not ready, you need to pass more blocks to add function.
+	struct Result
+	{
+		bool ready = false;
+		Block block;
+
+		Result(bool ready_) : ready(ready_) {}
+		Result(Block && block_) : ready(true), block(std::move(block_)) {}
+	};
+
+	/** Add next block and possibly returns squashed block.
+	  * At end, you need to pass empty block. As the result for last (empty) block, you will get last Result with ready = true.
+	  */
+	Result add(Block && block);
+
+private:
+	size_t min_block_size_rows;
+	size_t min_block_size_bytes;
+
+	Block accumulated_block;
+
+	void append(Block && block);
+
+	bool isEnoughSize(size_t rows, size_t bytes) const;
+};
+
+}

dbms/include/DB/IO/ReadHelpers.h

@@ -39,6 +39,8 @@ inline char parseEscapeSequence(char c)
 {
 	switch(c)
 	{
+		case 'a':
+			return '\a';
 		case 'b':
 			return '\b';
 		case 'f':
@@ -49,6 +51,8 @@ inline char parseEscapeSequence(char c)
 			return '\r';
 		case 't':
 			return '\t';
+		case 'v':
+			return '\v';
 		case '0':
 			return '\0';
 		default:

dbms/include/DB/Interpreters/Settings.h

@@ -39,6 +39,10 @@ struct Settings
 	M(SettingUInt64, max_block_size, DEFAULT_BLOCK_SIZE) \
 	/** Максимальный размер блока для вставки, если мы управляем формированием блоков для вставки. */ \
 	M(SettingUInt64, max_insert_block_size, DEFAULT_INSERT_BLOCK_SIZE) \
+	/** Squash blocks passed to INSERT query to specified size in rows, if blocks are not big enough. */ \
+	M(SettingUInt64, min_insert_block_size_rows, DEFAULT_INSERT_BLOCK_SIZE) \
+	/** Squash blocks passed to INSERT query to specified size in bytes, if blocks are not big enough. */ \
+	M(SettingUInt64, min_insert_block_size_bytes, (DEFAULT_INSERT_BLOCK_SIZE * 256)) \
 	/** Максимальное количество потоков выполнения запроса. По-умолчанию - определять автоматически. */ \
 	M(SettingMaxThreads, max_threads, 0) \
 	/** Максимальный размер буфера для чтения из файловой системы. */ \

dbms/src/Client/Client.cpp

@@ -1143,6 +1143,7 @@ public:
 			("time,t",			"print query execution time to stderr in non-interactive mode (for benchmarks)")
 			("stacktrace",		"print stack traces of exceptions")
 			("progress",		"print progress even in non-interactive mode")
+			("compression",		boost::program_options::value<bool>(),			"enable or disable compression")
 			APPLY_FOR_SETTINGS(DECLARE_SETTING)
 			APPLY_FOR_LIMITS(DECLARE_LIMIT)
 		;
@@ -1265,6 +1266,8 @@ public:
 			config().setBool("progress", true);
 		if (options.count("time"))
 			print_time_to_stderr = true;
+		if (options.count("compression"))
+			config().setBool("compression", options["compression"].as<bool>());
 	}
 };
 

dbms/src/Common/getNumberOfPhysicalCPUCores.cpp

@@ -1,4 +1,5 @@
 #include <DB/Common/getNumberOfPhysicalCPUCores.h>
+#include <thread>
 
 #if defined(__x86_64__)
 
@@ -7,10 +8,6 @@
 
 	namespace DB { namespace ErrorCodes { extern const int CPUID_ERROR; }}
 
-#elif defined(__aarch64__)
-
-	#include <thread>
-
 #endif
 
 
@@ -26,10 +23,14 @@ unsigned getNumberOfPhysicalCPUCores()
 	if (0 != cpu_identify(&raw_data, &data))
 		throw DB::Exception("Cannot cpu_identify: " + std::string(cpuid_error()), DB::ErrorCodes::CPUID_ERROR);
 
+	/// On Xen VMs, libcpuid returns wrong info (zero number of cores). Fallback to alternative method.
+	if (data.num_cores == 0 || data.total_logical_cpus == 0 || data.num_logical_cpus == 0)
+		return std::thread::hardware_concurrency();
+
 	return data.num_cores * data.total_logical_cpus / data.num_logical_cpus;
 
 #elif defined(__aarch64__)
-	/// Считаем, что на этой системе нет hyper-threading.
+	/// Assuming there are no hyper-threading on the system.
 	return std::thread::hardware_concurrency();
 #endif
 }

dbms/src/DataStreams/SquashingBlockInputStream.cpp

@@ -0,0 +1,31 @@
+#include <DB/DataStreams/SquashingBlockInputStream.h>
+
+
+namespace DB
+{
+
+SquashingBlockInputStream::SquashingBlockInputStream(BlockInputStreamPtr & src, size_t min_block_size_rows, size_t min_block_size_bytes)
+	: transform(min_block_size_rows, min_block_size_bytes)
+{
+	children.emplace_back(src);
+}
+
+
+Block SquashingBlockInputStream::readImpl()
+{
+	if (all_read)
+		return {};
+
+	while (true)
+	{
+		Block block = children[0]->read();
+		if (!block)
+			all_read = true;
+
+		SquashingTransform::Result result = transform.add(std::move(block));
+		if (result.ready)
+			return result.block;
+	}
+}
+
+}

dbms/src/DataStreams/SquashingBlockOutputStream.cpp

@@ -0,0 +1,53 @@
+#include <DB/DataStreams/SquashingBlockOutputStream.h>
+
+
+namespace DB
+{
+
+SquashingBlockOutputStream::SquashingBlockOutputStream(BlockOutputStreamPtr & dst, size_t min_block_size_rows, size_t min_block_size_bytes)
+	: output(dst), transform(min_block_size_rows, min_block_size_bytes)
+{
+}
+
+
+void SquashingBlockOutputStream::write(const Block & block)
+{
+	SquashingTransform::Result result = transform.add(Block(block));
+	if (result.ready)
+		output->write(result.block);
+}
+
+
+void SquashingBlockOutputStream::finalize()
+{
+	if (all_written)
+		return;
+
+	all_written = true;
+
+	SquashingTransform::Result result = transform.add({});
+	if (result.ready && result.block)
+		output->write(result.block);
+}
+
+
+void SquashingBlockOutputStream::flush()
+{
+	finalize();
+	output->flush();
+}
+
+
+void SquashingBlockOutputStream::writePrefix()
+{
+	output->writePrefix();
+}
+
+
+void SquashingBlockOutputStream::writeSuffix()
+{
+	finalize();
+	output->writeSuffix();
+}
+
+}

dbms/src/DataStreams/SquashingTransform.cpp

@@ -0,0 +1,76 @@
+#include <DB/DataStreams/SquashingTransform.h>
+
+
+namespace DB
+{
+
+SquashingTransform::SquashingTransform(size_t min_block_size_rows, size_t min_block_size_bytes)
+	: min_block_size_rows(min_block_size_rows), min_block_size_bytes(min_block_size_bytes)
+{
+}
+
+
+SquashingTransform::Result SquashingTransform::add(Block && block)
+{
+	if (!block)
+		return Result(std::move(accumulated_block));
+
+	/// Just read block is alredy enough.
+	if (isEnoughSize(block.rowsInFirstColumn(), block.bytes()))
+	{
+		/// If no accumulated data, return just read block.
+		if (!accumulated_block)
+			return Result(std::move(block));
+
+		/// Return accumulated data (may be it has small size) and place new block to accumulated data.
+		accumulated_block.swap(block);
+		return Result(std::move(block));
+	}
+
+	/// Accumulated block is already enough.
+	if (accumulated_block && isEnoughSize(accumulated_block.rowsInFirstColumn(), accumulated_block.bytes()))
+	{
+		/// Return accumulated data and place new block to accumulated data.
+		accumulated_block.swap(block);
+		return Result(std::move(block));
+	}
+
+	append(std::move(block));
+
+	if (isEnoughSize(accumulated_block.rowsInFirstColumn(), accumulated_block.bytes()))
+	{
+		Block res;
+		res.swap(accumulated_block);
+		return Result(std::move(res));
+	}
+
+	/// Squashed block is not ready.
+	return false;
+}
+
+
+void SquashingTransform::append(Block && block)
+{
+	if (!accumulated_block)
+	{
+		accumulated_block = std::move(block);
+		return;
+	}
+
+	size_t columns = block.columns();
+	size_t rows = block.rowsInFirstColumn();
+
+	for (size_t i = 0; i < columns; ++i)
+		accumulated_block.unsafeGetByPosition(i).column->insertRangeFrom(
+			*block.unsafeGetByPosition(i).column, 0, rows);
+}
+
+
+bool SquashingTransform::isEnoughSize(size_t rows, size_t bytes) const
+{
+	return (!min_block_size_rows && !min_block_size_bytes)
+		|| (min_block_size_rows && rows >= min_block_size_rows)
+		|| (min_block_size_bytes && bytes >= min_block_size_bytes);
+}
+
+}

dbms/src/Interpreters/InterpreterInsertQuery.cpp

@@ -5,6 +5,7 @@
 #include <DB/DataStreams/AddingDefaultBlockOutputStream.h>
 #include <DB/DataStreams/PushingToViewsBlockOutputStream.h>
 #include <DB/DataStreams/NullAndDoCopyBlockInputStream.h>
+#include <DB/DataStreams/SquashingBlockOutputStream.h>
 #include <DB/DataStreams/copyData.h>
 
 #include <DB/Parsers/ASTInsertQuery.h>
@@ -78,14 +79,21 @@ BlockIO InterpreterInsertQuery::execute()
 
 	NamesAndTypesListPtr required_columns = std::make_shared<NamesAndTypesList>(table->getColumnsList());
 
-	/// Создаем кортеж из нескольких стримов, в которые будем писать данные.
+	/// Создаем конвейер из нескольких стримов, в которые будем писать данные.
-	BlockOutputStreamPtr out =
+	BlockOutputStreamPtr out;
-		std::make_shared<ProhibitColumnsBlockOutputStream>(
+
-			std::make_shared<AddingDefaultBlockOutputStream>(
+	out = std::make_shared<PushingToViewsBlockOutputStream>(query.database, query.table, context, query_ptr);
-				std::make_shared<MaterializingBlockOutputStream>(
+
-					std::make_shared<PushingToViewsBlockOutputStream>(query.database, query.table, context, query_ptr)),
+	out = std::make_shared<MaterializingBlockOutputStream>(out);
-				required_columns, table->column_defaults, context, static_cast<bool>(context.getSettingsRef().strict_insert_defaults)),
+
-			table->materialized_columns);
+	out = std::make_shared<AddingDefaultBlockOutputStream>(out,
+		required_columns, table->column_defaults, context, static_cast<bool>(context.getSettingsRef().strict_insert_defaults));
+
+	out = std::make_shared<ProhibitColumnsBlockOutputStream>(out, table->materialized_columns);
+
+	out = std::make_shared<SquashingBlockOutputStream>(out,
+		context.getSettingsRef().min_insert_block_size_rows,
+		context.getSettingsRef().min_insert_block_size_bytes);
 
 	BlockIO res;
 	res.out_sample = getSampleBlock();
@@ -98,7 +106,8 @@ BlockIO InterpreterInsertQuery::execute()
 	else
 	{
 		InterpreterSelectQuery interpreter_select{query.select, context};
-		BlockInputStreamPtr in{interpreter_select.execute().in};
+		BlockInputStreamPtr in = interpreter_select.execute().in;
+
 		res.in = std::make_shared<NullAndDoCopyBlockInputStream>(in, out);
 		res.in_sample = interpreter_select.getSampleBlock();
 	}

dbms/src/Interpreters/InterpreterSelectQuery.cpp

@@ -785,6 +785,9 @@ QueryProcessingStage::Enum InterpreterSelectQuery::executeFetchColumns()
 	{
 		size_t max_streams = settings.max_threads;
 
+		if (max_streams == 0)
+			throw Exception("Logical error: zero number of streams requested", ErrorCodes::LOGICAL_ERROR);
+
 		/// Если надо - запрашиваем больше источников, чем количество потоков - для более равномерного распределения работы по потокам.
 		if (max_streams > 1 && !is_remote)
 			max_streams *= settings.max_streams_to_max_threads_ratio;

dbms/src/Server/TCPHandler.cpp

@@ -626,7 +626,7 @@ void TCPHandler::initBlockOutput()
 	{
 		if (state.compression == Protocol::Compression::Enable)
 			state.maybe_compressed_out = std::make_shared<CompressedWriteBuffer>(
-				*out, query_context.getSettings().network_compression_method);
+				*out, query_context.getSettingsRef().network_compression_method);
 		else
 			state.maybe_compressed_out = out;
 

dbms/src/Server/TCPHandler.h

@@ -18,26 +18,26 @@ namespace DB
 {
 
 
-/// Состояние обработки запроса.
+/// State of query processing.
 struct QueryState
 {
-	/// Идентификатор запроса.
+	/// Identifier of the query.
 	String query_id;
 
 	QueryProcessingStage::Enum stage = QueryProcessingStage::Complete;
 	Protocol::Compression::Enum compression = Protocol::Compression::Disable;
 
-	/// Откуда читать данные для INSERT-а.
+	/// From where to read data for INSERT.
 	std::shared_ptr<ReadBuffer> maybe_compressed_in;
 	BlockInputStreamPtr block_in;
 
-	/// Куда писать возвращаемые данные.
+	/// Where to write result data.
 	std::shared_ptr<WriteBuffer> maybe_compressed_out;
 	BlockOutputStreamPtr block_out;
 
-	/// Текст запроса.
+	/// Query text.
 	String query;
-	/// Потоки блоков, с помощью которых выполнять запрос.
+	/// Streams of blocks, that are processing the query.
 	BlockIO io;
 
 	/// Отменен ли запрос

dbms/tests/queries/0_stateless/00325_replacing_merge_tree.sql

@@ -57,9 +57,7 @@ INSERT INTO test.replacing VALUES ('2000-01-01', 2, 'def', 1);
 INSERT INTO test.replacing VALUES ('2000-01-01', 2, 'ghi', 0);
 SELECT * FROM test.replacing FINAL ORDER BY k, v, _part_index;
 
-OPTIMIZE TABLE test.replacing;
+OPTIMIZE TABLE test.replacing PARTITION 200001 FINAL;
-OPTIMIZE TABLE test.replacing;
-OPTIMIZE TABLE test.replacing;
 SELECT _part_index, * FROM test.replacing ORDER BY k, v, _part_index;
 SELECT _part_index, * FROM test.replacing FINAL ORDER BY k, v, _part_index;
 
@@ -91,9 +89,7 @@ INSERT INTO test.replacing VALUES ('2000-01-01', 2, 'def');
 INSERT INTO test.replacing VALUES ('2000-01-01', 2, 'ghi');
 SELECT * FROM test.replacing FINAL ORDER BY k, _part_index;
 
-OPTIMIZE TABLE test.replacing;
+OPTIMIZE TABLE test.replacing PARTITION 200001 FINAL;
-OPTIMIZE TABLE test.replacing;
-OPTIMIZE TABLE test.replacing;
 SELECT _part_index, * FROM test.replacing ORDER BY k, _part_index;
 SELECT _part_index, * FROM test.replacing FINAL ORDER BY k, _part_index;
 

dbms/tests/queries/0_stateless/00340_squashing_insert_select.reference

@@ -0,0 +1,5 @@
+1000000	10
+130000	76
+1000000	10
+120000	1
+20000000

dbms/tests/queries/0_stateless/00340_squashing_insert_select.sql

@@ -0,0 +1,18 @@
+DROP TABLE IF EXISTS test.numbers_squashed;
+CREATE TABLE test.numbers_squashed AS system.numbers ENGINE = StripeLog;
+
+SET max_block_size = 10000;
+
+SET min_insert_block_size_rows = 1000000;
+SET min_insert_block_size_bytes = 0;
+
+INSERT INTO test.numbers_squashed SELECT * FROM system.numbers LIMIT 10000000;
+SELECT blockSize() AS b, count() / b AS c FROM test.numbers_squashed GROUP BY blockSize() ORDER BY c DESC;
+
+SET min_insert_block_size_bytes = 1000000;
+INSERT INTO test.numbers_squashed SELECT * FROM system.numbers LIMIT 10000000;
+SELECT blockSize() AS b, count() / b AS c FROM test.numbers_squashed GROUP BY blockSize() ORDER BY c DESC;
+
+SELECT count() FROM test.numbers_squashed;
+
+DROP TABLE test.numbers_squashed;

dbms/tests/queries/0_stateless/00341_squashing_insert_select2.reference

@@ -0,0 +1,17 @@
+10	2
+100	1
+120
+10	3
+100	3
+330
+100	5
+10	4
+540
+100	6
+10	5
+30	1
+680
+10	9
+100	7
+30	1
+820

dbms/tests/queries/0_stateless/00341_squashing_insert_select2.sql

@@ -0,0 +1,68 @@
+DROP TABLE IF EXISTS test.numbers_squashed;
+CREATE TABLE test.numbers_squashed (number UInt8) ENGINE = Memory;
+
+SET min_insert_block_size_rows = 100;
+SET min_insert_block_size_bytes = 0;
+SET max_threads = 1;
+
+INSERT INTO test.numbers_squashed
+SELECT arrayJoin(range(10)) AS number
+UNION ALL
+SELECT arrayJoin(range(100))
+UNION ALL
+SELECT arrayJoin(range(10));
+
+SELECT blockSize() AS b, count() / b AS c FROM test.numbers_squashed GROUP BY blockSize() ORDER BY c DESC, b ASC;
+SELECT count() FROM test.numbers_squashed;
+
+INSERT INTO test.numbers_squashed
+SELECT arrayJoin(range(100)) AS number
+UNION ALL
+SELECT arrayJoin(range(10))
+UNION ALL
+SELECT arrayJoin(range(100));
+
+SELECT blockSize() AS b, count() / b AS c FROM test.numbers_squashed GROUP BY blockSize() ORDER BY c DESC, b ASC;
+SELECT count() FROM test.numbers_squashed;
+
+INSERT INTO test.numbers_squashed
+SELECT arrayJoin(range(10)) AS number
+UNION ALL
+SELECT arrayJoin(range(100))
+UNION ALL
+SELECT arrayJoin(range(100));
+
+SELECT blockSize() AS b, count() / b AS c FROM test.numbers_squashed GROUP BY blockSize() ORDER BY c DESC, b ASC;
+SELECT count() FROM test.numbers_squashed;
+
+INSERT INTO test.numbers_squashed
+SELECT arrayJoin(range(10)) AS number
+UNION ALL
+SELECT arrayJoin(range(10))
+UNION ALL
+SELECT arrayJoin(range(10))
+UNION ALL
+SELECT arrayJoin(range(100))
+UNION ALL
+SELECT arrayJoin(range(10));
+
+SELECT blockSize() AS b, count() / b AS c FROM test.numbers_squashed GROUP BY blockSize() ORDER BY c DESC, b ASC;
+SELECT count() FROM test.numbers_squashed;
+
+SET min_insert_block_size_rows = 10;
+
+INSERT INTO test.numbers_squashed
+SELECT arrayJoin(range(10)) AS number
+UNION ALL
+SELECT arrayJoin(range(10))
+UNION ALL
+SELECT arrayJoin(range(10))
+UNION ALL
+SELECT arrayJoin(range(100))
+UNION ALL
+SELECT arrayJoin(range(10));
+
+SELECT blockSize() AS b, count() / b AS c FROM test.numbers_squashed GROUP BY blockSize() ORDER BY c DESC, b ASC;
+SELECT count() FROM test.numbers_squashed;
+
+DROP TABLE test.numbers_squashed;

dbms/tests/queries/0_stateless/00342_escape_sequences.reference

@@ -0,0 +1 @@
+07080C0A0D090B5C27223FAA

dbms/tests/queries/0_stateless/00342_escape_sequences.sql

@@ -0,0 +1 @@
+SELECT hex('\a\b\f\n\r\t\v\\\'\"\?\xAA');

doc/index.html

@@ -399,8 +399,10 @@
 
 		<code>
 <pre>
+sudo apt-key adv --keyserver keyserver.ubuntu.com --recv E0C56BD4    # optional
+
 sudo mkdir -p /etc/apt/sources.list.d
-echo "deb http://repo.yandex.ru/clickhouse/<span id="distributive">trusty</span>/ dists/stable/main/binary-amd64/" |
+echo "deb http://repo.yandex.ru/clickhouse/<span id="distributive">trusty</span> stable main" |
 &nbsp;&nbsp;&nbsp;&nbsp;sudo tee /etc/apt/sources.list.d/clickhouse.list
 sudo apt-get update
 

doc/reference_en.html

@@ -417,23 +417,26 @@ In %%/etc/apt/sources.list%% (or in a separate %%/etc/apt/sources.list.d/clickho
 On Ubuntu Trusty (14.04):
 
 %%
-deb http://repo.yandex.ru/clickhouse/trusty/ dists/stable/main/binary-amd64/
+deb http://repo.yandex.ru/clickhouse/trusty stable main
 %%
 
 On Ubuntu Precise (12.04):
 
 %%
-deb http://repo.yandex.ru/clickhouse/precise/ dists/stable/main/binary-amd64/
+deb http://repo.yandex.ru/clickhouse/precise stable main
 %%
 
 Then run:
 
 %%
+sudo apt-key adv --keyserver keyserver.ubuntu.com --recv E0C56BD4    # optional
 sudo apt-get update
 sudo apt-get install clickhouse-client clickhouse-server-common
 %%
 
-You can also download and install packages manually from here: <a href="https://dist.yandex.ru/metrika/stable/amd64/">https://dist.yandex.ru/metrika/stable/amd64/</a>.
+You can also download and install packages manually from here:
+<a href="http://repo.yandex.ru/clickhouse/trusty/pool/main/c/clickhouse/">http://repo.yandex.ru/clickhouse/trusty/pool/main/c/clickhouse/</a>,
+<a href="http://repo.yandex.ru/clickhouse/precise/pool/main/c/clickhouse/">http://repo.yandex.ru/clickhouse/precise/pool/main/c/clickhouse/</a>.
 
 ClickHouse contains access restriction settings. They are located in the &#39;users.xml&#39; file (next to &#39;config.xml&#39;).
 By default, access is allowed from everywhere for the default user without a password. See &#39;user/default/networks&#39;. For more information, see the section &quot;Configuration files&quot;.
@@ -2810,7 +2813,7 @@ Replication is an optional feature. To use replication, set the addresses of the
 &lt;/zookeeper>
 %%
 
-Use ZooKeeper version 3.4.5 or later. For example, the version in the Ubuntu Precise package is too old. You can get a newer version for Ubuntu Precise from the repository <a href="https://dist.yandex.net/metrika/stable/amd64/">https://dist.yandex.net/metrika/stable/amd64/</a>.
+Use ZooKeeper version 3.4.5 or later. For example, the version in the Ubuntu Precise package is too old.
 
 You can specify any existing ZooKeeper cluster - the system will use a directory on it for its own data (the directory is specified when creating a replicatable table).
 

doc/reference_ru.html

@@ -424,25 +424,27 @@ YT позволяет хранить группы столбцов по отде
 На Ubuntu Trusty (14.04):
 
 %%
-deb http://repo.yandex.ru/clickhouse/trusty/ dists/stable/main/binary-amd64/
+deb http://repo.yandex.ru/clickhouse/trusty stable main
 %%
 
 На Ubuntu Precise (12.04):
 
 %%
-deb http://repo.yandex.ru/clickhouse/precise/ dists/stable/main/binary-amd64/
+deb http://repo.yandex.ru/clickhouse/precise stable main
 %%
 
 
 Затем выполните:
 
 %%
+sudo apt-key adv --keyserver keyserver.ubuntu.com --recv E0C56BD4    # optional
 sudo apt-get update
 sudo apt-get install clickhouse-client clickhouse-server-common
 %%
 
 Также можно скачать и установить пакеты вручную, отсюда:
-<a href="https://dist.yandex.ru/metrika/stable/amd64/">https://dist.yandex.ru/metrika/stable/amd64/</a>.
+<a href="http://repo.yandex.ru/clickhouse/trusty/pool/main/c/clickhouse/">http://repo.yandex.ru/clickhouse/trusty/pool/main/c/clickhouse/</a>,
+<a href="http://repo.yandex.ru/clickhouse/precise/pool/main/c/clickhouse/">http://repo.yandex.ru/clickhouse/precise/pool/main/c/clickhouse/</a>.
 
 ClickHouse содержит настройки ограничения доступа. Они расположены в файле users.xml (рядом с config.xml).
 По умолчанию, разрешён доступ отовсюду для пользователя default без пароля. См. секцию users/default/networks.
@@ -2867,7 +2869,7 @@ min_bytes, max_bytes - условие на количество байт в бу
 &lt;/zookeeper&gt;
 %%
 
-Используйте версию ZooKeeper не старее 3.4.5. Для примера, в Ubuntu Precise слишком старая версия в пакете. Достаточно новую версию для Ubuntu Precise можно взять из репозитория <a href='https://dist.yandex.net/metrika/stable/amd64/'>https://dist.yandex.net/metrika/stable/amd64/</a>.
+Используйте версию ZooKeeper не старее 3.4.5. Для примера, в Ubuntu Precise слишком старая версия в пакете.
 
 Можно указать любой имеющийся у вас ZooKeeper-кластер - система будет использовать в нём одну директорию для своих данных (директория указывается при создании реплицируемой таблицы).