thevar1able/ClickHouse
1
0
Fork 0
mirror of https://github.com/ClickHouse/ClickHouse.git synced 2024-09-21 09:10:48 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge

Browse Source
This commit is contained in:
Evgeniy Gatov 2015-10-19 23:26:57 +03:00
commit caf2d7fd1e
5 changed files with 214 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

138
dbms/include/DB/Functions/FunctionsStringArray.h
View File

@ -414,8 +414,142 @@ public:
};
typedef FunctionTokens<AlphaTokensImpl> FunctionAlphaTokens;
typedef FunctionTokens<SplitByCharImpl> FunctionSplitByChar;
/// Склеивает массив строк в одну строку через разделитель.
class FunctionArrayStringConcat : public IFunction
{
private:
void executeImpl(
const ColumnString::Chars_t & src_chars,
const ColumnString::Offsets_t & src_string_offsets,
const ColumnArray::Offsets_t & src_array_offsets,
const char * delimiter, const size_t delimiter_size,
ColumnString::Chars_t & dst_chars,
ColumnString::Offsets_t & dst_string_offsets)
{
size_t size = src_array_offsets.size();
if (!size)
return;
/// С небольшим запасом - как будто разделитель идёт и после последней строки массива.
dst_chars.resize(
src_chars.size()
+ delimiter_size * src_string_offsets.size() /// Разделители после каждой строки...
+ src_array_offsets.size() /// Нулевой байт после каждой склеенной строки
- src_string_offsets.size()); /// Бывший нулевой байт после каждой строки массива
/// Будет столько строк, сколько было массивов.
dst_string_offsets.resize(src_array_offsets.size());
ColumnArray::Offset_t current_src_array_offset = 0;
ColumnString::Offset_t current_src_string_offset = 0;
ColumnString::Offset_t current_dst_string_offset = 0;
/// Цикл по массивам строк.
for (size_t i = 0; i < size; ++i)
{
/// Цикл по строкам внутри массива. /// NOTE Можно всё сделать за одно копирование, если разделитель имеет размер 1.
for (auto next_src_array_offset = src_array_offsets[i]; current_src_array_offset < next_src_array_offset; ++current_src_array_offset)
{
size_t bytes_to_copy = src_string_offsets[current_src_array_offset] - current_src_string_offset - 1;
memcpy(&dst_chars[current_dst_string_offset], &src_chars[current_src_string_offset], bytes_to_copy);
current_src_string_offset = src_string_offsets[current_src_array_offset];
current_dst_string_offset += bytes_to_copy;
if (current_src_array_offset + 1 != next_src_array_offset)
{
memcpy(&dst_chars[current_dst_string_offset], delimiter, delimiter_size);
current_dst_string_offset += delimiter_size;
}
}
dst_chars[current_dst_string_offset] = 0;
++current_dst_string_offset;
dst_string_offsets[i] = current_dst_string_offset;
}
dst_chars.resize(dst_string_offsets.back());
}
public:
static constexpr auto name = "arrayStringConcat";
static IFunction * create(const Context & context) { return new FunctionArrayStringConcat; }
/// Получить имя функции.
String getName() const override
{
return name;
}
/// Получить тип результата по типам аргументов. Если функция неприменима для данных аргументов - кинуть исключение.
DataTypePtr getReturnType(const DataTypes & arguments) const override
{
if (arguments.size() != 1 && arguments.size() != 2)
throw Exception("Number of arguments for function " + getName() + " doesn't match: passed "
+ toString(arguments.size()) + ", should be 1 or 2.",
ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH);
const DataTypeArray * array_type = typeid_cast<const DataTypeArray *>(arguments[0].get());
if (!array_type || !typeid_cast<const DataTypeString *>(array_type->getNestedType().get()))
throw Exception("First argument for function " + getName() + " must be array of strings.", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
if (arguments.size() == 2
&& !typeid_cast<const DataTypeString *>(arguments[1].get()))
throw Exception("Second argument for function " + getName() + " must be constant string.", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
return new DataTypeString;
}
/// Выполнить функцию над блоком.
void execute(Block & block, const ColumnNumbers & arguments, size_t result) override
{
String delimiter;
if (arguments.size() == 2)
{
const ColumnConstString * col_delim = typeid_cast<const ColumnConstString *>(block.getByPosition(arguments[1]).column.get());
if (!col_delim)
throw Exception("Second argument for function " + getName() + " must be constant string.", ErrorCodes::ILLEGAL_COLUMN);
delimiter = col_delim->getData();
}
if (const ColumnConstArray * col_const_arr = typeid_cast<const ColumnConstArray *>(block.getByPosition(arguments[0]).column.get()))
{
ColumnConstString * col_res = new ColumnConstString(col_const_arr->size(), "");
block.getByPosition(result).column = col_res;
const Array & src_arr = col_const_arr->getData();
String & dst_str = col_res->getData();
for (size_t i = 0, size = src_arr.size(); i < size; ++i)
{
if (i != 0)
dst_str += delimiter;
dst_str += src_arr[i].get<const String &>();
}
}
else
{
const ColumnArray & col_arr = static_cast<const ColumnArray &>(*block.getByPosition(arguments[0]).column);
const ColumnString & col_string = static_cast<const ColumnString &>(col_arr.getData());
ColumnString * col_res = new ColumnString;
block.getByPosition(result).column = col_res;
executeImpl(
col_string.getChars(), col_string.getOffsets(), col_arr.getOffsets(),
delimiter.data(), delimiter.size(),
col_res->getChars(), col_res->getOffsets());
}
}
};
typedef FunctionTokens<AlphaTokensImpl> FunctionAlphaTokens;
typedef FunctionTokens<SplitByCharImpl> FunctionSplitByChar;
typedef FunctionTokens<SplitByStringImpl> FunctionSplitByString;
typedef FunctionTokens<ExtractAllImpl> FunctionExtractAll;

1
dbms/src/Functions/FunctionsStringArray.cpp
View File

@ -10,6 +10,7 @@ void registerFunctionsStringArray(FunctionFactory & factory)
factory.registerFunction<FunctionAlphaTokens>();
factory.registerFunction<FunctionSplitByChar>();
factory.registerFunction<FunctionSplitByString>();
factory.registerFunction<FunctionArrayStringConcat>();
}
}

3
dbms/tests/queries/0_stateless/00058_global_in.sh
View File

@ -25,7 +25,8 @@ cat "$CONFIG" | sed -r \
s/<http_port>[0-9]+/<http_port>8124/;
s/<tcp_port>[0-9]+/<tcp_port>9001/;
s/<interserver_http_port>[0-9]+/<interserver_http_port>9010/;
s/<use_olap_http_server>true/<use_olap_http_server>false/;
s/<olap_compatibility>/<!-- <olap_compatibility>/;
s/<\/olap_compatibility>/<\/olap_compatibility> -->/;
s/users\.xml/users-preprocessed.xml/' > $CONFIG2
cp ${CONFIG/config/users} .

65
dbms/tests/queries/0_stateless/00255_array_concat_string.reference Normal file
View File

@ -0,0 +1,65 @@
HelloWorld
HelloWorld
Hello, World
Hello, World
0
01
012
0123
01234
012345
0123456
01234567
012345678
0
01
012
0123
01234
012345
0123456
01234567
012345678
0
0,1
0,1,2
0,1,2,3
0,1,2,3,4
0,1,2,3,4,5
0,1,2,3,4,5,6
0,1,2,3,4,5,6,7
0,1,2,3,4,5,6,7,8
yandex
yandex google
yandex google test
yandex google test 123
yandex google test 123
yandex google test 123 hello
yandex google test 123 hello world
yandex google test 123 hello world goodbye
yandex google test 123 hello world goodbye xyz
yandex google test 123 hello world goodbye xyz yandex
yandex google test 123 hello world goodbye xyz yandex google
yandex google test 123 hello world goodbye xyz yandex google test
yandex google test 123 hello world goodbye xyz yandex google test 123
yandex google test 123 hello world goodbye xyz yandex google test 123
yandex google test 123 hello world goodbye xyz yandex google test 123 hello
yandex google test 123 hello world goodbye xyz yandex google test 123 hello world
yandex google test 123 hello world goodbye xyz yandex google test 123 hello world goodbye
yandex google test 123 hello world goodbye xyz yandex google test 123 hello world goodbye xyz
yandex google test 123 hello world goodbye xyz yandex google test 123 hello world goodbye xyz yandex
0
01
012
0
01
012
0

10
dbms/tests/queries/0_stateless/00255_array_concat_string.sql Normal file
View File

@ -0,0 +1,10 @@
SELECT arrayStringConcat(['Hello', 'World']);
SELECT arrayStringConcat(materialize(['Hello', 'World']));
SELECT arrayStringConcat(['Hello', 'World'], ', ');
SELECT arrayStringConcat(materialize(['Hello', 'World']), ', ');
SELECT arrayStringConcat(emptyArrayString());
SELECT arrayStringConcat(arrayMap(x -> toString(x), range(number))) FROM system.numbers LIMIT 10;
SELECT arrayStringConcat(arrayMap(x -> toString(x), range(number)), '') FROM system.numbers LIMIT 10;
SELECT arrayStringConcat(arrayMap(x -> toString(x), range(number)), ',') FROM system.numbers LIMIT 10;
SELECT arrayStringConcat(arrayMap(x -> transform(x, [0, 1, 2, 3, 4, 5, 6, 7, 8], ['yandex', 'google', 'test', '123', '', 'hello', 'world', 'goodbye', 'xyz'], ''), arrayMap(x -> x % 9, range(number))), ' ') FROM system.numbers LIMIT 20;
SELECT arrayStringConcat(arrayMap(x -> toString(x), range(number % 4))) FROM system.numbers LIMIT 10;