ClickHouse/dbms/src/Storages/StorageHDFS.cpp

241 lines
6.8 KiB
C++
Raw Normal View History

#include <Common/config.h>
#if USE_HDFS
2018-11-19 08:17:09 +00:00
#include <Storages/StorageFactory.h>
#include <Storages/StorageHDFS.h>
2018-11-19 08:17:09 +00:00
#include <Interpreters/Context.h>
#include <Interpreters/evaluateConstantExpression.h>
#include <Parsers/ASTLiteral.h>
#include <IO/ReadBufferFromHDFS.h>
#include <IO/WriteBufferFromHDFS.h>
2019-07-24 09:51:02 +00:00
#include <IO/HDFSCommon.h>
2018-11-19 08:17:09 +00:00
#include <Formats/FormatFactory.h>
#include <DataStreams/IBlockOutputStream.h>
#include <DataStreams/UnionBlockInputStream.h>
#include <DataStreams/IBlockInputStream.h>
2018-11-19 08:17:09 +00:00
#include <DataStreams/OwningBlockInputStream.h>
2019-07-24 09:51:02 +00:00
#include <Common/parseGlobs.h>
#include <Poco/URI.h>
#include <re2/re2.h>
#include <re2/stringpiece.h>
#include <hdfs/hdfs.h>
2018-11-19 08:17:09 +00:00
namespace DB
{
namespace ErrorCodes
{
extern const int NUMBER_OF_ARGUMENTS_DOESNT_MATCH;
2018-11-25 04:22:01 +00:00
extern const int NOT_IMPLEMENTED;
extern const int BAD_ARGUMENTS;
2018-11-19 08:17:09 +00:00
}
StorageHDFS::StorageHDFS(const String & uri_,
const std::string & database_name_,
2018-11-19 08:17:09 +00:00
const std::string & table_name_,
const String & format_name_,
const ColumnsDescription & columns_,
Context & context_)
: IStorage(columns_)
, uri(uri_)
, format_name(format_name_)
, table_name(table_name_)
, database_name(database_name_)
, context(context_)
2018-11-19 08:17:09 +00:00
{
}
namespace
{
class HDFSBlockInputStream : public IBlockInputStream
{
public:
HDFSBlockInputStream(const String & uri,
const String & format,
const Block & sample_block,
const Context & context,
2019-02-10 16:55:12 +00:00
UInt64 max_block_size)
{
std::unique_ptr<ReadBuffer> read_buf = std::make_unique<ReadBufferFromHDFS>(uri);
auto input_stream = FormatFactory::instance().getInput(format, *read_buf, sample_block, context, max_block_size);
reader = std::make_shared<OwningBlockInputStream<ReadBuffer>>(input_stream, std::move(read_buf));
}
String getName() const override
2018-11-19 08:17:09 +00:00
{
return "HDFS";
}
Block readImpl() override
{
return reader->read();
}
Block getHeader() const override
{
return reader->getHeader();
}
void readPrefixImpl() override
{
reader->readPrefix();
}
void readSuffixImpl() override
{
reader->readSuffix();
}
private:
BlockInputStreamPtr reader;
};
class HDFSBlockOutputStream : public IBlockOutputStream
{
public:
HDFSBlockOutputStream(const String & uri,
const String & format,
const Block & sample_block_,
const Context & context)
: sample_block(sample_block_)
{
write_buf = std::make_unique<WriteBufferFromHDFS>(uri);
writer = FormatFactory::instance().getOutput(format, *write_buf, sample_block, context);
}
Block getHeader() const override
{
return sample_block;
}
void write(const Block & block) override
{
writer->write(block);
}
void writePrefix() override
{
writer->writePrefix();
}
void writeSuffix() override
{
writer->writeSuffix();
writer->flush();
write_buf->sync();
}
private:
Block sample_block;
std::unique_ptr<WriteBufferFromHDFS> write_buf;
BlockOutputStreamPtr writer;
};
2018-11-19 08:17:09 +00:00
2019-07-24 09:51:02 +00:00
//static Strings recursiveLSWithRegexpMatching(const String & cur_path, hdfsFS fs, const re2::RE2 & matcher)
//{
// HDFSFileInfo ls;
// ls.file_info = hdfsListDirectory(fs.get(), path_without_globs, ls.length);
//
//}
2018-11-19 08:17:09 +00:00
}
BlockInputStreams StorageHDFS::read(
const Names & /*column_names*/,
const SelectQueryInfo & /*query_info*/,
2019-01-17 16:15:51 +00:00
const Context & context_,
2018-11-19 08:17:09 +00:00
QueryProcessingStage::Enum /*processed_stage*/,
size_t max_block_size,
2018-11-19 08:17:09 +00:00
unsigned /*num_streams*/)
{
2019-07-24 09:51:02 +00:00
Strings path_parts;
2019-07-30 18:46:49 +00:00
size_t first_glob = uri.find_first_of("*?{");
if (first_glob == std::string::npos)
2019-07-24 09:51:02 +00:00
return {std::make_shared<HDFSBlockInputStream>(
uri,
format_name,
getSampleBlock(),
context_,
max_block_size)};
2019-07-30 18:46:49 +00:00
String uri_without_globs = uri.substr(0, first_glob);
size_t end_of_path_without_globs = uri_without_globs.rfind('/');
uri_without_globs = uri_without_globs.substr(0, end_of_path_without_globs + 1);
size_t begin_of_path = uri.find('/', uri.find("//") + 2);
String path_from_uri = uri.substr(begin_of_path);
String uri_without_path = uri.substr(0, begin_of_path);
String path_pattern = makeRegexpPatternFromGlobs(path_from_uri);
2019-07-24 09:51:02 +00:00
re2::RE2 matcher(path_pattern);
2019-07-30 18:46:49 +00:00
HDFSBuilderPtr builder = createHDFSBuilder(Poco::URI(uri_without_globs));
2019-07-24 09:51:02 +00:00
HDFSFSPtr fs = createHDFSFS(builder.get());
// Strings res_paths = recursiveLSWithRegexpMatching(path_without_globs, fs.get(), matcher);
2019-07-30 18:46:49 +00:00
2019-07-24 09:51:02 +00:00
HDFSFileInfo ls;
2019-07-30 18:46:49 +00:00
String path_without_globs = uri_without_globs.substr(begin_of_path);
2019-07-24 09:51:02 +00:00
ls.file_info = hdfsListDirectory(fs.get(), path_without_globs.data(), &ls.length);
BlockInputStreams result;
for (int i = 0; i < ls.length; ++i)
{
if (ls.file_info[i].mKind == 'F')
{
2019-07-30 18:46:49 +00:00
String cur_path = String(ls.file_info[i].mName);
if (cur_path[1] == '/')
{
cur_path = cur_path.substr(1);
}
2019-07-24 09:51:02 +00:00
if (re2::RE2::FullMatch(cur_path, matcher))
{
result.push_back(
2019-07-30 18:46:49 +00:00
std::make_shared<HDFSBlockInputStream>(uri_without_path + String(ls.file_info[i].mName), format_name, getSampleBlock(), context_,
2019-07-24 09:51:02 +00:00
max_block_size));
}
}
}
return result;
2018-11-19 08:17:09 +00:00
}
void StorageHDFS::rename(const String & /*new_path_to_db*/, const String & new_database_name, const String & new_table_name)
{
table_name = new_table_name;
database_name = new_database_name;
}
2018-11-19 08:17:09 +00:00
BlockOutputStreamPtr StorageHDFS::write(const ASTPtr & /*query*/, const Context & /*context*/)
2018-11-19 08:17:09 +00:00
{
return std::make_shared<HDFSBlockOutputStream>(uri, format_name, getSampleBlock(), context);
2018-11-19 08:17:09 +00:00
}
void registerStorageHDFS(StorageFactory & factory)
{
factory.registerStorage("HDFS", [](const StorageFactory::Arguments & args)
{
ASTs & engine_args = args.engine_args;
if (!(engine_args.size() == 1 || engine_args.size() == 2))
throw Exception(
"Storage HDFS requires exactly 2 arguments: url and name of used format.", ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH);
engine_args[0] = evaluateConstantExpressionOrIdentifierAsLiteral(engine_args[0], args.local_context);
2019-03-15 17:09:14 +00:00
String url = engine_args[0]->as<ASTLiteral &>().value.safeGet<String>();
2018-11-19 08:17:09 +00:00
engine_args[1] = evaluateConstantExpressionOrIdentifierAsLiteral(engine_args[1], args.local_context);
2019-03-15 17:09:14 +00:00
String format_name = engine_args[1]->as<ASTLiteral &>().value.safeGet<String>();
2018-11-19 08:17:09 +00:00
return StorageHDFS::create(url, args.database_name, args.table_name, format_name, args.columns, args.context);
2018-11-19 08:17:09 +00:00
});
}
}
#endif