ClickHouse/dbms/include/DB/Functions/FunctionsString.h

1174 lines
38 KiB
C
Raw Normal View History

2011-10-03 04:06:34 +00:00
#pragma once
2011-10-03 05:29:11 +00:00
#include <Poco/UTF8Encoding.h>
#include <Poco/Unicode.h>
2011-10-03 04:06:34 +00:00
#include <DB/DataTypes/DataTypesNumberFixed.h>
#include <DB/DataTypes/DataTypeString.h>
#include <DB/DataTypes/DataTypeFixedString.h>
#include <DB/DataTypes/DataTypeArray.h>
2011-10-03 04:06:34 +00:00
#include <DB/Columns/ColumnString.h>
#include <DB/Columns/ColumnArray.h>
2011-10-03 04:06:34 +00:00
#include <DB/Columns/ColumnFixedString.h>
#include <DB/Columns/ColumnConst.h>
#include <DB/Functions/IFunction.h>
namespace DB
{
/** Функции работы со строками:
*
* length, empty, notEmpty,
* concat, substring, lower, upper, reverse, // left, right, insert, replace,
* // escape, quote, unescape, unquote, trim, trimLeft, trimRight, pad, padLeft
* lengthUTF8, substringUTF8, lowerUTF8, upperUTF8, reverseUTF8, // leftUTF8, rightUTF8, insertUTF8, replaceUTF8
* // padUTF8, padLeftUTF8.
2011-10-03 04:06:34 +00:00
*
* s -> UInt8: empty, notEmpty
2011-10-03 04:06:34 +00:00
* s -> UInt64: length, lengthUTF8
2011-10-10 06:10:49 +00:00
* s -> s: lower, upper, lowerUTF8, upperUTF8, reverse, reverseUTF8, escape, quote, unescape, unquote, trim, trimLeft, trimRight
2011-10-03 04:06:34 +00:00
* s, s -> s: concat
2011-10-10 10:05:39 +00:00
* s, c1, c2 -> s: substring, substringUTF8
* s, c1 -> s: left, right, leftUTF8, rightUTF8
* s, c1, s2 -> s: insert, insertUTF8, pad, padLeft, padUTF8, padLeftUTF8
2011-10-03 04:06:34 +00:00
* s, c1, c2, s2 -> s: replace, replaceUTF8
*
2011-10-03 04:06:34 +00:00
* Функции поиска строк и регулярных выражений расположены отдельно.
* Функции работы с URL расположены отдельно.
* Функции кодирования строк, конвертации в другие типы расположены отдельно.
2012-09-03 04:45:46 +00:00
*
* Функции length, empty, notEmpty также работают с массивами.
2011-10-03 04:06:34 +00:00
*/
template <bool negative = false>
struct EmptyImpl
{
2013-09-15 05:51:43 +00:00
static void vector(const ColumnString::Chars_t & data, const ColumnString::Offsets_t & offsets,
PODArray<UInt8> & res)
{
size_t size = offsets.size();
ColumnString::Offset_t prev_offset = 1;
for (size_t i = 0; i < size; ++i)
{
res[i] = negative ^ (offsets[i] == prev_offset);
prev_offset = offsets[i] + 1;
}
}
2013-09-15 05:51:43 +00:00
static void vector_fixed_to_constant(const ColumnString::Chars_t & data, size_t n,
UInt8 & res)
{
res = negative ^ (n == 0);
}
2013-09-15 05:51:43 +00:00
static void vector_fixed_to_vector(const ColumnString::Chars_t & data, size_t n,
PODArray<UInt8> & res)
{
}
static void constant(const std::string & data, UInt8 & res)
{
res = negative ^ (data.empty());
}
static void array(const ColumnString::Offsets_t & offsets, PODArray<UInt8> & res)
{
size_t size = offsets.size();
ColumnString::Offset_t prev_offset = 0;
for (size_t i = 0; i < size; ++i)
{
res[i] = negative ^ (offsets[i] == prev_offset);
prev_offset = offsets[i];
}
}
static void constant_array(const Array & data, UInt8 & res)
{
res = negative ^ (data.empty());
}
};
2011-10-03 04:06:34 +00:00
/** Вычисляет длину строки в байтах.
*/
struct LengthImpl
{
2013-09-15 05:51:43 +00:00
static void vector(const ColumnString::Chars_t & data, const ColumnString::Offsets_t & offsets,
PODArray<UInt64> & res)
2011-10-03 04:06:34 +00:00
{
size_t size = offsets.size();
for (size_t i = 0; i < size; ++i)
res[i] = i == 0
? (offsets[i] - 1)
: (offsets[i] - 1 - offsets[i - 1]);
}
2013-09-15 05:51:43 +00:00
static void vector_fixed_to_constant(const ColumnString::Chars_t & data, size_t n,
2011-10-03 04:06:34 +00:00
UInt64 & res)
{
res = n;
}
2013-09-15 05:51:43 +00:00
static void vector_fixed_to_vector(const ColumnString::Chars_t & data, size_t n,
PODArray<UInt64> & res)
2011-10-03 04:06:34 +00:00
{
}
static void constant(const std::string & data, UInt64 & res)
{
res = data.size();
}
static void array(const ColumnString::Offsets_t & offsets, PODArray<UInt64> & res)
{
size_t size = offsets.size();
for (size_t i = 0; i < size; ++i)
res[i] = i == 0
? (offsets[i])
: (offsets[i] - offsets[i - 1]);
}
static void constant_array(const Array & data, UInt64 & res)
{
res = data.size();
}
2011-10-03 04:06:34 +00:00
};
/** Если строка представляет собой текст в кодировке UTF-8, то возвращает длину текста в кодовых точках.
* (не в символах: длина текста "ё" может быть как 1, так и 2, в зависимости от нормализации)
* Иначе - поведение не определено.
*/
struct LengthUTF8Impl
{
2013-09-15 05:51:43 +00:00
static void vector(const ColumnString::Chars_t & data, const ColumnString::Offsets_t & offsets,
PODArray<UInt64> & res)
2011-10-03 04:06:34 +00:00
{
size_t size = offsets.size();
ColumnString::Offset_t prev_offset = 0;
2011-10-03 04:06:34 +00:00
for (size_t i = 0; i < size; ++i)
{
res[i] = 0;
for (const UInt8 * c = &data[prev_offset]; c + 1 < &data[offsets[i]]; ++c)
if (*c <= 0x7F || *c >= 0xC0)
++res[i];
prev_offset = offsets[i];
}
}
2013-09-15 05:51:43 +00:00
static void vector_fixed_to_constant(const ColumnString::Chars_t & data, size_t n,
2011-10-03 04:06:34 +00:00
UInt64 & res)
{
}
2013-09-15 05:51:43 +00:00
static void vector_fixed_to_vector(const ColumnString::Chars_t & data, size_t n,
PODArray<UInt64> & res)
2011-10-03 04:06:34 +00:00
{
size_t size = data.size() / n;
for (size_t i = 0; i < size; ++i)
{
res[i] = 0;
for (const UInt8 * c = &data[i * n]; c < &data[(i + 1) * n]; ++c)
if (*c <= 0x7F || *c >= 0xC0)
++res[i];
}
}
static void constant(const std::string & data, UInt64 & res)
{
res = 0;
for (const UInt8 * c = reinterpret_cast<const UInt8 *>(data.data()); c < reinterpret_cast<const UInt8 *>(data.data() + data.size()); ++c)
if (*c <= 0x7F || *c >= 0xC0)
++res;
}
static void array(const ColumnString::Offsets_t & offsets, PODArray<UInt64> & res)
{
throw Exception("Cannot apply function lengthUTF8 to Array argument", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
}
static void constant_array(const Array & data, UInt64 & res)
{
throw Exception("Cannot apply function lengthUTF8 to Array argument", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
}
2011-10-03 04:06:34 +00:00
};
2011-10-03 05:29:11 +00:00
/** Переводит строку в нижний (верхний) регистр, в текущей локали, в однобайтовой кодировке.
*/
template <int F(int)>
struct LowerUpperImpl
{
2013-09-15 05:51:43 +00:00
static void vector(const ColumnString::Chars_t & data, const ColumnString::Offsets_t & offsets,
ColumnString::Chars_t & res_data, ColumnString::Offsets_t & res_offsets)
2011-10-03 05:29:11 +00:00
{
res_data.resize(data.size());
res_offsets.assign(offsets);
2011-10-03 05:29:11 +00:00
array(&*data.begin(), &*data.end(), &*res_data.begin());
}
2013-09-15 05:51:43 +00:00
static void vector_fixed(const ColumnString::Chars_t & data, size_t n,
ColumnString::Chars_t & res_data)
2011-10-03 05:29:11 +00:00
{
res_data.resize(data.size());
array(&*data.begin(), &*data.end(), &*res_data.begin());
}
static void constant(const std::string & data, std::string & res_data)
{
res_data.resize(data.size());
array(reinterpret_cast<const UInt8 *>(&*data.begin()), reinterpret_cast<const UInt8 *>(&*data.end()),
reinterpret_cast<UInt8 *>(&*res_data.begin()));
}
private:
static void array(const UInt8 * src, const UInt8 * src_end, UInt8 * dst)
{
for (; src < src_end; ++src, ++dst)
*dst = F(*src);
}
};
/** Если строка содержит текст в кодировке UTF-8 - перевести его в нижний (верхний) регистр.
* Замечание: предполагается, что после перевода символа в другой регистр,
* длина его мультибайтовой последовательности в UTF-8 не меняется.
* Иначе - поведение не определено.
*/
template <int F(int)>
struct LowerUpperUTF8Impl
{
2013-09-15 05:51:43 +00:00
static void vector(const ColumnString::Chars_t & data, const ColumnString::Offsets_t & offsets,
ColumnString::Chars_t & res_data, ColumnString::Offsets_t & res_offsets)
2011-10-03 05:29:11 +00:00
{
res_data.resize(data.size());
res_offsets.assign(offsets);
2011-10-03 05:29:11 +00:00
array(&*data.begin(), &*data.end(), &*res_data.begin());
}
2013-09-15 05:51:43 +00:00
static void vector_fixed(const ColumnString::Chars_t & data, size_t n,
ColumnString::Chars_t & res_data)
2011-10-03 05:29:11 +00:00
{
res_data.resize(data.size());
array(&*data.begin(), &*data.end(), &*res_data.begin());
}
static void constant(const std::string & data, std::string & res_data)
{
res_data.resize(data.size());
array(reinterpret_cast<const UInt8 *>(&*data.begin()), reinterpret_cast<const UInt8 *>(&*data.end()),
reinterpret_cast<UInt8 *>(&*res_data.begin()));
}
private:
static void array(const UInt8 * src, const UInt8 * src_end, UInt8 * dst)
{
static Poco::UTF8Encoding utf8;
2011-10-03 05:29:11 +00:00
while (src < src_end)
{
int chars = utf8.convert(F(utf8.convert(src)), dst, src_end - src);
if (chars)
{
src += chars;
dst += chars;
}
else
{
++src;
++dst;
}
}
}
};
2011-10-10 06:10:49 +00:00
/** Разворачивает строку в байтах.
*/
struct ReverseImpl
{
2013-09-15 05:51:43 +00:00
static void vector(const ColumnString::Chars_t & data, const ColumnString::Offsets_t & offsets,
ColumnString::Chars_t & res_data, ColumnString::Offsets_t & res_offsets)
2011-10-10 06:10:49 +00:00
{
res_data.resize(data.size());
res_offsets.assign(offsets);
2011-10-10 06:10:49 +00:00
size_t size = offsets.size();
ColumnString::Offset_t prev_offset = 0;
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
{
for (size_t j = prev_offset; j < offsets[i] - 1; ++j)
res_data[j] = data[offsets[i] + prev_offset - 2 - j];
res_data[offsets[i] - 1] = 0;
prev_offset = offsets[i];
}
}
2013-09-15 05:51:43 +00:00
static void vector_fixed(const ColumnString::Chars_t & data, size_t n,
ColumnString::Chars_t & res_data)
2011-10-10 06:10:49 +00:00
{
res_data.resize(data.size());
size_t size = data.size() / n;
for (size_t i = 0; i < size; ++i)
for (size_t j = i * n; j < (i + 1) * n; ++j)
res_data[j] = data[(i * 2 + 1) * n - j - 1];
}
static void constant(const std::string & data, std::string & res_data)
{
res_data.resize(data.size());
for (size_t j = 0; j < data.size(); ++j)
res_data[j] = data[data.size() - j - 1];
}
};
/** Разворачивает последовательность кодовых точек в строке в кодировке UTF-8.
* Результат может не соответствовать ожидаемому, так как модифицирующие кодовые точки (например, диакритика) могут примениться не к тем символам.
* Если строка не в кодировке UTF-8, то поведение не определено.
*/
struct ReverseUTF8Impl
{
2013-09-15 05:51:43 +00:00
static void vector(const ColumnString::Chars_t & data, const ColumnString::Offsets_t & offsets,
ColumnString::Chars_t & res_data, ColumnString::Offsets_t & res_offsets)
2011-10-10 06:10:49 +00:00
{
res_data.resize(data.size());
res_offsets.assign(offsets);
2011-10-10 06:10:49 +00:00
size_t size = offsets.size();
ColumnString::Offset_t prev_offset = 0;
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
{
ColumnString::Offset_t j = prev_offset;
2011-10-10 06:10:49 +00:00
while (j < offsets[i] - 1)
{
if (data[j] < 0xBF)
{
res_data[offsets[i] + prev_offset - 2 - j] = data[j];
j += 1;
}
else if (data[j] < 0xE0)
{
memcpy(&res_data[offsets[i] + prev_offset - 2 - j - 1], &data[j], 2);
j += 2;
}
else if (data[j] < 0xF0)
{
memcpy(&res_data[offsets[i] + prev_offset - 2 - j - 2], &data[j], 3);
j += 3;
}
else
{
res_data[offsets[i] + prev_offset - 2 - j] = data[j];
j += 1;
}
}
2011-10-10 06:10:49 +00:00
res_data[offsets[i] - 1] = 0;
prev_offset = offsets[i];
}
}
2013-09-15 05:51:43 +00:00
static void vector_fixed(const ColumnString::Chars_t & data, size_t n,
ColumnString::Chars_t & res_data)
2011-10-10 06:10:49 +00:00
{
throw Exception("Cannot apply function reverseUTF8 to fixed string.", ErrorCodes::ILLEGAL_COLUMN);
}
static void constant(const std::string & data, std::string & res_data)
{
res_data.resize(data.size());
size_t j = 0;
while (j < data.size())
{
if (static_cast<unsigned char>(data[j]) < 0xBF)
{
res_data[data.size() - 1 - j] = data[j];
j += 1;
}
else if (static_cast<unsigned char>(data[j]) < 0xE0)
{
memcpy(&res_data[data.size() - 1 - j - 1], &data[j], 2);
j += 2;
}
else if (static_cast<unsigned char>(data[j]) < 0xF0)
{
memcpy(&res_data[data.size() - 1 - j - 2], &data[j], 3);
j += 3;
}
else
{
res_data[data.size() - 1 - j] = data[j];
j += 1;
}
}
}
};
/** Склеивает две строки.
*/
struct ConcatImpl
{
static void vector_vector(
2013-09-15 05:51:43 +00:00
const ColumnString::Chars_t & a_data, const ColumnString::Offsets_t & a_offsets,
const ColumnString::Chars_t & b_data, const ColumnString::Offsets_t & b_offsets,
ColumnString::Chars_t & c_data, ColumnString::Offsets_t & c_offsets)
2011-10-10 06:10:49 +00:00
{
size_t size = a_offsets.size();
c_data.resize(a_data.size() + b_data.size() - size);
c_offsets.resize(size);
ColumnString::Offset_t offset = 0;
ColumnString::Offset_t a_offset = 0;
ColumnString::Offset_t b_offset = 0;
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
{
memcpy(&c_data[offset], &a_data[a_offset], a_offsets[i] - a_offset - 1);
offset += a_offsets[i] - a_offset - 1;
memcpy(&c_data[offset], &b_data[b_offset], b_offsets[i] - b_offset);
offset += b_offsets[i] - b_offset;
2011-10-10 06:10:49 +00:00
a_offset = a_offsets[i];
b_offset = b_offsets[i];
2011-10-10 06:10:49 +00:00
c_offsets[i] = offset;
}
}
static void vector_fixed_vector(
2013-09-15 05:51:43 +00:00
const ColumnString::Chars_t & a_data, const ColumnString::Offsets_t & a_offsets,
const ColumnString::Chars_t & b_data, ColumnString::Offset_t b_n,
ColumnString::Chars_t & c_data, ColumnString::Offsets_t & c_offsets)
2011-10-10 06:10:49 +00:00
{
size_t size = a_offsets.size();
c_data.resize(a_data.size() + b_data.size());
c_offsets.resize(size);
ColumnString::Offset_t offset = 0;
ColumnString::Offset_t a_offset = 0;
ColumnString::Offset_t b_offset = 0;
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
{
memcpy(&c_data[offset], &a_data[a_offset], a_offsets[i] - a_offset - 1);
offset += a_offsets[i] - a_offset - 1;
memcpy(&c_data[offset], &b_data[b_offset], b_n);
offset += b_n;
c_data[offset] = 0;
offset += 1;
a_offset = a_offsets[i];
b_offset += b_n;
c_offsets[i] = offset;
}
}
static void vector_constant(
2013-09-15 05:51:43 +00:00
const ColumnString::Chars_t & a_data, const ColumnString::Offsets_t & a_offsets,
2011-10-10 06:10:49 +00:00
const std::string & b,
2013-09-15 05:51:43 +00:00
ColumnString::Chars_t & c_data, ColumnString::Offsets_t & c_offsets)
2011-10-10 06:10:49 +00:00
{
size_t size = a_offsets.size();
c_data.resize(a_data.size() + b.size() * size);
c_offsets.assign(a_offsets);
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
c_offsets[i] += b.size() * (i + 1);
2011-10-10 06:10:49 +00:00
ColumnString::Offset_t offset = 0;
ColumnString::Offset_t a_offset = 0;
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
{
memcpy(&c_data[offset], &a_data[a_offset], a_offsets[i] - a_offset - 1);
offset += a_offsets[i] - a_offset - 1;
memcpy(&c_data[offset], b.data(), b.size() + 1);
offset += b.size() + 1;
a_offset = a_offsets[i];
}
}
static void fixed_vector_vector(
2013-09-15 05:51:43 +00:00
const ColumnString::Chars_t & a_data, ColumnString::Offset_t a_n,
const ColumnString::Chars_t & b_data, const ColumnString::Offsets_t & b_offsets,
ColumnString::Chars_t & c_data, ColumnString::Offsets_t & c_offsets)
2011-10-10 06:10:49 +00:00
{
size_t size = b_offsets.size();
c_data.resize(a_data.size() + b_data.size());
c_offsets.resize(size);
ColumnString::Offset_t offset = 0;
ColumnString::Offset_t a_offset = 0;
ColumnString::Offset_t b_offset = 0;
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
{
memcpy(&c_data[offset], &a_data[a_offset], a_n);
offset += a_n;
memcpy(&c_data[offset], &b_data[b_offset], b_offsets[i] - b_offset);
offset += b_offsets[i] - b_offset;
a_offset = a_n;
b_offset = b_offsets[i];
2011-10-10 06:10:49 +00:00
c_offsets[i] = offset;
}
}
static void fixed_vector_fixed_vector(
2013-09-15 05:51:43 +00:00
const ColumnString::Chars_t & a_data, ColumnString::Offset_t a_n,
const ColumnString::Chars_t & b_data, ColumnString::Offset_t b_n,
ColumnString::Chars_t & c_data, ColumnString::Offsets_t & c_offsets)
2011-10-10 06:10:49 +00:00
{
size_t size = a_data.size() / a_n;
c_data.resize(a_data.size() + b_data.size() + size);
c_offsets.resize(size);
2011-10-10 06:10:49 +00:00
ColumnString::Offset_t offset = 0;
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
{
memcpy(&c_data[offset], &a_data[i * a_n], a_n);
offset += a_n;
memcpy(&c_data[offset], &b_data[i * b_n], b_n);
offset += b_n;
c_data[offset] = 0;
++offset;
c_offsets[i] = offset;
2011-10-10 06:10:49 +00:00
}
}
static void fixed_vector_constant(
2013-09-15 05:51:43 +00:00
const ColumnString::Chars_t & a_data, ColumnString::Offset_t a_n,
2011-10-10 06:10:49 +00:00
const std::string & b,
2013-09-15 05:51:43 +00:00
ColumnString::Chars_t & c_data, ColumnString::Offsets_t & c_offsets)
2011-10-10 06:10:49 +00:00
{
size_t size = a_data.size() / a_n;
ColumnString::Offset_t b_n = b.size();
c_data.resize(a_data.size() + size * b_n + size);
c_offsets.resize(size);
2011-10-10 06:10:49 +00:00
ColumnString::Offset_t offset = 0;
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
{
memcpy(&c_data[offset], &a_data[i * a_n], a_n);
offset += a_n;
memcpy(&c_data[offset], b.data(), b_n);
offset += b_n;
c_data[offset] = 0;
++offset;
c_offsets[i] = offset;
2011-10-10 06:10:49 +00:00
}
}
static void constant_vector(
const std::string & a,
2013-09-15 05:51:43 +00:00
const ColumnString::Chars_t & b_data, const ColumnString::Offsets_t & b_offsets,
ColumnString::Chars_t & c_data, ColumnString::Offsets_t & c_offsets)
2011-10-10 06:10:49 +00:00
{
size_t size = b_offsets.size();
c_data.resize(b_data.size() + a.size() * size);
c_offsets.assign(b_offsets);
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
c_offsets[i] += a.size() * (i + 1);
2011-10-10 06:10:49 +00:00
ColumnString::Offset_t offset = 0;
ColumnString::Offset_t b_offset = 0;
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
{
memcpy(&c_data[offset], a.data(), a.size());
offset += a.size();
memcpy(&c_data[offset], &b_data[b_offset], b_offsets[i] - b_offset);
offset += b_offsets[i] - b_offset;
b_offset = b_offsets[i];
}
}
static void constant_fixed_vector(
const std::string & a,
2013-09-15 05:51:43 +00:00
const ColumnString::Chars_t & b_data, ColumnString::Offset_t b_n,
ColumnString::Chars_t & c_data, ColumnString::Offsets_t & c_offsets)
2011-10-10 06:10:49 +00:00
{
size_t size = b_data.size() / b_n;
ColumnString::Offset_t a_n = a.size();
c_data.resize(size * a_n + b_data.size() + size);
c_offsets.resize(size);
2011-10-10 06:10:49 +00:00
ColumnString::Offset_t offset = 0;
2011-10-10 06:10:49 +00:00
for (size_t i = 0; i < size; ++i)
{
memcpy(&c_data[offset], a.data(), a_n);
offset += a_n;
memcpy(&c_data[offset], &b_data[i * b_n], b_n);
offset += b_n;
c_data[offset] = 0;
++offset;
c_offsets[i] = offset;
2011-10-10 06:10:49 +00:00
}
}
static void constant_constant(
const std::string & a,
const std::string & b,
std::string & c)
{
c = a + b;
}
};
2011-10-10 10:05:39 +00:00
/** Выделяет подстроку в строке, как последовательности байт.
*/
struct SubstringImpl
{
2013-09-15 05:51:43 +00:00
static void vector(const ColumnString::Chars_t & data, const ColumnString::Offsets_t & offsets,
2011-10-10 10:05:39 +00:00
size_t start, size_t length,
2013-09-15 05:51:43 +00:00
ColumnString::Chars_t & res_data, ColumnString::Offsets_t & res_offsets)
2011-10-10 10:05:39 +00:00
{
res_data.reserve(data.size());
size_t size = offsets.size();
res_offsets.resize(size);
ColumnString::Offset_t prev_offset = 0;
ColumnString::Offset_t res_offset = 0;
2011-10-10 10:05:39 +00:00
for (size_t i = 0; i < size; ++i)
{
if (start + prev_offset >= offsets[i] + 1)
{
res_data.resize(res_data.size() + 1);
res_data[res_offset] = 0;
++res_offset;
}
else
{
size_t bytes_to_copy = std::min(offsets[i] - prev_offset - start, length);
res_data.resize(res_data.size() + bytes_to_copy + 1);
memcpy(&res_data[res_offset], &data[prev_offset + start - 1], bytes_to_copy);
res_offset += bytes_to_copy + 1;
res_data[res_offset - 1] = 0;
}
res_offsets[i] = res_offset;
prev_offset = offsets[i];
}
}
2013-09-15 05:51:43 +00:00
static void vector_fixed(const ColumnString::Chars_t & data, size_t n,
2011-10-10 10:05:39 +00:00
size_t start, size_t length,
2013-09-15 05:51:43 +00:00
ColumnString::Chars_t & res_data, ColumnString::Offsets_t & res_offsets)
2011-10-10 10:05:39 +00:00
{
if (length == 0 || start + length > n + 1)
throw Exception("Index out of bound for function substring of fixed size value", ErrorCodes::ARGUMENT_OUT_OF_BOUND);
2011-10-10 10:05:39 +00:00
size_t size = data.size() / n;
res_offsets.resize(size);
res_data.resize(length * size + size);
2011-10-10 10:05:39 +00:00
ColumnString::Offset_t res_offset = 0;
2011-10-10 10:05:39 +00:00
for (size_t i = 0; i < size; ++i)
{
memcpy(&res_data[res_offset], &data[i * n + start - 1], length);
res_offset += length;
res_data[res_offset] = 0;
++res_offset;
res_offsets[i] = res_offset;
}
2011-10-10 10:05:39 +00:00
}
static void constant(const std::string & data,
size_t start, size_t length,
std::string & res_data)
{
if (start + length > data.size() + 1)
throw Exception("Index out of bound for function substring of fixed size value", ErrorCodes::ARGUMENT_OUT_OF_BOUND);
2011-10-10 10:05:39 +00:00
res_data = data.substr(start - 1, length);
}
};
2011-10-10 12:26:26 +00:00
/** Если строка в кодировке UTF-8, то выделяет в ней подстроку кодовых точек.
* Иначе - поведение не определено.
*/
struct SubstringUTF8Impl
{
2013-09-15 05:51:43 +00:00
static void vector(const ColumnString::Chars_t & data, const ColumnString::Offsets_t & offsets,
2011-10-10 12:26:26 +00:00
size_t start, size_t length,
2013-09-15 05:51:43 +00:00
ColumnString::Chars_t & res_data, ColumnString::Offsets_t & res_offsets)
2011-10-10 12:26:26 +00:00
{
res_data.reserve(data.size());
size_t size = offsets.size();
res_offsets.resize(size);
ColumnString::Offset_t prev_offset = 0;
ColumnString::Offset_t res_offset = 0;
2011-10-10 12:26:26 +00:00
for (size_t i = 0; i < size; ++i)
{
ColumnString::Offset_t j = prev_offset;
ColumnString::Offset_t pos = 1;
ColumnString::Offset_t bytes_start = 0;
ColumnString::Offset_t bytes_length = 0;
2011-10-10 12:26:26 +00:00
while (j < offsets[i] - 1)
{
if (pos == start)
bytes_start = j - prev_offset + 1;
if (data[j] < 0xBF)
j += 1;
else if (data[j] < 0xE0)
j += 2;
else if (data[j] < 0xF0)
j += 3;
else
j += 1;
2011-10-10 12:26:26 +00:00
if (pos >= start && pos < start + length)
bytes_length = j - prev_offset + 1 - bytes_start;
else if (pos >= start + length)
break;
++pos;
}
if (bytes_start == 0)
{
res_data.resize(res_data.size() + 1);
res_data[res_offset] = 0;
++res_offset;
}
else
{
size_t bytes_to_copy = std::min(offsets[i] - prev_offset - bytes_start, bytes_length);
res_data.resize(res_data.size() + bytes_to_copy + 1);
memcpy(&res_data[res_offset], &data[prev_offset + bytes_start - 1], bytes_to_copy);
res_offset += bytes_to_copy + 1;
res_data[res_offset - 1] = 0;
}
res_offsets[i] = res_offset;
prev_offset = offsets[i];
}
}
2013-09-15 05:51:43 +00:00
static void vector_fixed(const ColumnString::Chars_t & data, ColumnString::Offset_t n,
2011-10-10 12:26:26 +00:00
size_t start, size_t length,
2013-09-15 05:51:43 +00:00
ColumnString::Chars_t & res_data, ColumnString::Offsets_t & res_offsets)
2011-10-10 12:26:26 +00:00
{
throw Exception("Cannot apply function substringUTF8 to fixed string.", ErrorCodes::ILLEGAL_COLUMN);
}
static void constant(const std::string & data,
size_t start, size_t length,
std::string & res_data)
{
if (start + length > data.size() + 1)
throw Exception("Index out of bound for function substring of constant value", ErrorCodes::ARGUMENT_OUT_OF_BOUND);
ColumnString::Offset_t j = 0;
ColumnString::Offset_t pos = 1;
ColumnString::Offset_t bytes_start = 0;
ColumnString::Offset_t bytes_length = 0;
2011-10-10 12:26:26 +00:00
while (j < data.size())
{
if (pos == start)
bytes_start = j + 1;
if (static_cast<unsigned char>(data[j]) < 0xBF)
j += 1;
else if (static_cast<unsigned char>(data[j]) < 0xE0)
j += 2;
else if (static_cast<unsigned char>(data[j]) < 0xF0)
j += 3;
else
j += 1;
if (pos >= start && pos < start + length)
bytes_length = j + 1 - bytes_start;
else if (pos >= start + length)
break;
2011-10-10 12:26:26 +00:00
++pos;
}
if (bytes_start != 0)
res_data = data.substr(bytes_start - 1, bytes_length);
}
};
template <typename Impl, typename Name, typename ResultType>
class FunctionStringOrArrayToT : public IFunction
2011-10-03 04:06:34 +00:00
{
public:
/// Получить имя функции.
String getName() const
{
return Name::get();
}
/// Получить тип результата по типам аргументов. Если функция неприменима для данных аргументов - кинуть исключение.
DataTypePtr getReturnType(const DataTypes & arguments) const
{
if (arguments.size() != 1)
throw Exception("Number of arguments for function " + getName() + " doesn't match: passed "
2013-06-21 20:34:19 +00:00
+ toString(arguments.size()) + ", should be 1.",
2011-10-03 04:06:34 +00:00
ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH);
if (!typeid_cast<const DataTypeString *>(&*arguments[0]) && !typeid_cast<const DataTypeFixedString *>(&*arguments[0])
&& !typeid_cast<const DataTypeArray *>(&*arguments[0]))
2011-10-03 04:06:34 +00:00
throw Exception("Illegal type " + arguments[0]->getName() + " of argument of function " + getName(),
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
return new typename DataTypeFromFieldType<ResultType>::Type;
2011-10-03 04:06:34 +00:00
}
/// Выполнить функцию над блоком.
void execute(Block & block, const ColumnNumbers & arguments, size_t result)
{
const ColumnPtr column = block.getByPosition(arguments[0]).column;
if (const ColumnString * col = typeid_cast<const ColumnString *>(&*column))
2011-10-03 04:06:34 +00:00
{
ColumnVector<ResultType> * col_res = new ColumnVector<ResultType>;
block.getByPosition(result).column = col_res;
typename ColumnVector<ResultType>::Container_t & vec_res = col_res->getData();
2011-10-03 04:06:34 +00:00
vec_res.resize(col->size());
Impl::vector(col->getChars(), col->getOffsets(), vec_res);
2011-10-03 04:06:34 +00:00
}
else if (const ColumnFixedString * col = typeid_cast<const ColumnFixedString *>(&*column))
2011-10-03 04:06:34 +00:00
{
/// Для фиксированной строки, только функция lengthUTF8 возвращает не константу.
if ("lengthUTF8" != getName())
2011-10-03 04:06:34 +00:00
{
ResultType res = 0;
Impl::vector_fixed_to_constant(col->getChars(), col->getN(), res);
2011-10-03 04:06:34 +00:00
ColumnConst<ResultType> * col_res = new ColumnConst<ResultType>(col->size(), res);
block.getByPosition(result).column = col_res;
}
else
{
ColumnVector<ResultType> * col_res = new ColumnVector<ResultType>;
block.getByPosition(result).column = col_res;
typename ColumnVector<ResultType>::Container_t & vec_res = col_res->getData();
2011-10-03 04:06:34 +00:00
vec_res.resize(col->size());
Impl::vector_fixed_to_vector(col->getChars(), col->getN(), vec_res);
2011-10-03 04:06:34 +00:00
}
}
else if (const ColumnConstString * col = typeid_cast<const ColumnConstString *>(&*column))
2011-10-03 04:06:34 +00:00
{
ResultType res = 0;
Impl::constant(col->getData(), res);
ColumnConst<ResultType> * col_res = new ColumnConst<ResultType>(col->size(), res);
block.getByPosition(result).column = col_res;
}
else if (const ColumnArray * col = typeid_cast<const ColumnArray *>(&*column))
{
ColumnVector<ResultType> * col_res = new ColumnVector<ResultType>;
block.getByPosition(result).column = col_res;
typename ColumnVector<ResultType>::Container_t & vec_res = col_res->getData();
vec_res.resize(col->size());
Impl::array(col->getOffsets(), vec_res);
}
else if (const ColumnConstArray * col = typeid_cast<const ColumnConstArray *>(&*column))
{
ResultType res = 0;
Impl::constant_array(col->getData(), res);
ColumnConst<ResultType> * col_res = new ColumnConst<ResultType>(col->size(), res);
block.getByPosition(result).column = col_res;
}
2011-10-03 04:06:34 +00:00
else
throw Exception("Illegal column " + block.getByPosition(arguments[0]).column->getName()
+ " of argument of function " + getName(),
ErrorCodes::ILLEGAL_COLUMN);
}
};
2011-10-03 05:29:11 +00:00
template <typename Impl, typename Name>
class FunctionStringToString : public IFunction
{
public:
/// Получить имя функции.
String getName() const
{
return Name::get();
}
/// Получить тип результата по типам аргументов. Если функция неприменима для данных аргументов - кинуть исключение.
DataTypePtr getReturnType(const DataTypes & arguments) const
{
if (arguments.size() != 1)
throw Exception("Number of arguments for function " + getName() + " doesn't match: passed "
2013-06-21 20:34:19 +00:00
+ toString(arguments.size()) + ", should be 1.",
2011-10-03 05:29:11 +00:00
ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH);
if (!typeid_cast<const DataTypeString *>(&*arguments[0]) && !typeid_cast<const DataTypeFixedString *>(&*arguments[0]))
2011-10-03 05:29:11 +00:00
throw Exception("Illegal type " + arguments[0]->getName() + " of argument of function " + getName(),
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
return arguments[0]->clone();
2011-10-03 05:29:11 +00:00
}
/// Выполнить функцию над блоком.
void execute(Block & block, const ColumnNumbers & arguments, size_t result)
{
const ColumnPtr column = block.getByPosition(arguments[0]).column;
if (const ColumnString * col = typeid_cast<const ColumnString *>(&*column))
2011-10-03 05:29:11 +00:00
{
ColumnString * col_res = new ColumnString;
block.getByPosition(result).column = col_res;
Impl::vector(col->getChars(), col->getOffsets(),
col_res->getChars(), col_res->getOffsets());
2011-10-03 05:29:11 +00:00
}
else if (const ColumnFixedString * col = typeid_cast<const ColumnFixedString *>(&*column))
2011-10-03 05:29:11 +00:00
{
ColumnFixedString * col_res = new ColumnFixedString(col->getN());
block.getByPosition(result).column = col_res;
Impl::vector_fixed(col->getChars(), col->getN(),
col_res->getChars());
2011-10-03 05:29:11 +00:00
}
else if (const ColumnConstString * col = typeid_cast<const ColumnConstString *>(&*column))
2011-10-03 05:29:11 +00:00
{
String res;
Impl::constant(col->getData(), res);
ColumnConstString * col_res = new ColumnConstString(col->size(), res);
block.getByPosition(result).column = col_res;
}
else
throw Exception("Illegal column " + block.getByPosition(arguments[0]).column->getName()
+ " of argument of function " + getName(),
ErrorCodes::ILLEGAL_COLUMN);
}
};
2011-10-10 06:10:49 +00:00
template <typename Impl, typename Name>
class FunctionStringStringToString : public IFunction
{
public:
/// Получить имя функции.
String getName() const
{
return Name::get();
}
/// Получить тип результата по типам аргументов. Если функция неприменима для данных аргументов - кинуть исключение.
DataTypePtr getReturnType(const DataTypes & arguments) const
{
if (arguments.size() != 2)
throw Exception("Number of arguments for function " + getName() + " doesn't match: passed "
2013-06-21 20:34:19 +00:00
+ toString(arguments.size()) + ", should be 2.",
2011-10-10 06:10:49 +00:00
ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH);
if (!typeid_cast<const DataTypeString *>(&*arguments[0]) && !typeid_cast<const DataTypeFixedString *>(&*arguments[0]))
2011-10-10 06:10:49 +00:00
throw Exception("Illegal type " + arguments[0]->getName() + " of first argument of function " + getName(),
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
if (!typeid_cast<const DataTypeString *>(&*arguments[1]) && !typeid_cast<const DataTypeFixedString *>(&*arguments[1]))
2011-10-10 06:10:49 +00:00
throw Exception("Illegal type " + arguments[1]->getName() + " of second argument of function " + getName(),
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
return new DataTypeString;
}
/// Выполнить функцию над блоком.
void execute(Block & block, const ColumnNumbers & arguments, size_t result)
{
const IColumn * c0 = &*block.getByPosition(arguments[0]).column;
const IColumn * c1 = &*block.getByPosition(arguments[1]).column;
2011-10-10 06:10:49 +00:00
const ColumnString * c0_string = typeid_cast<const ColumnString *>(c0);
const ColumnString * c1_string = typeid_cast<const ColumnString *>(c1);
const ColumnFixedString * c0_fixed_string = typeid_cast<const ColumnFixedString *>(c0);
const ColumnFixedString * c1_fixed_string = typeid_cast<const ColumnFixedString *>(c1);
const ColumnConstString * c0_const = typeid_cast<const ColumnConstString *>(c0);
const ColumnConstString * c1_const = typeid_cast<const ColumnConstString *>(c1);
2011-10-10 06:10:49 +00:00
/// Результат - const string
if (c0_const && c1_const)
{
ColumnConstString * c_res = new ColumnConstString(c0_const->size(), "");
block.getByPosition(result).column = c_res;
Impl::constant_constant(c0_const->getData(), c1_const->getData(), c_res->getData());
}
else
{
ColumnString * c_res = new ColumnString;
block.getByPosition(result).column = c_res;
ColumnString::Chars_t & vec_res = c_res->getChars();
ColumnString::Offsets_t & offsets_res = c_res->getOffsets();
if (c0_string && c1_string)
Impl::vector_vector(
c0_string->getChars(), c0_string->getOffsets(),
c1_string->getChars(), c1_string->getOffsets(),
vec_res, offsets_res);
else if (c0_string && c1_fixed_string)
Impl::vector_fixed_vector(
c0_string->getChars(), c0_string->getOffsets(),
c1_fixed_string->getChars(), c1_fixed_string->getN(),
vec_res, offsets_res);
else if (c0_string && c1_const)
Impl::vector_constant(
c0_string->getChars(), c0_string->getOffsets(),
c1_const->getData(),
vec_res, offsets_res);
else if (c0_fixed_string && c1_string)
Impl::fixed_vector_vector(
c0_fixed_string->getChars(), c0_fixed_string->getN(),
c1_string->getChars(), c1_string->getOffsets(),
vec_res, offsets_res);
else if (c0_const && c1_string)
Impl::constant_vector(
c0_const->getData(),
c1_string->getChars(), c1_string->getOffsets(),
vec_res, offsets_res);
else if (c0_fixed_string && c1_fixed_string)
Impl::fixed_vector_fixed_vector(
c0_fixed_string->getChars(), c0_fixed_string->getN(),
c1_fixed_string->getChars(), c1_fixed_string->getN(),
vec_res, offsets_res);
else if (c0_fixed_string && c1_const)
Impl::fixed_vector_constant(
c0_fixed_string->getChars(), c0_fixed_string->getN(),
c1_const->getData(),
vec_res, offsets_res);
else if (c0_const && c1_fixed_string)
Impl::constant_fixed_vector(
c0_const->getData(),
c1_fixed_string->getChars(), c1_fixed_string->getN(),
vec_res, offsets_res);
else
throw Exception("Illegal columns "
+ block.getByPosition(arguments[0]).column->getName() + " and "
+ block.getByPosition(arguments[1]).column->getName()
+ " of arguments of function " + getName(),
ErrorCodes::ILLEGAL_COLUMN);
2011-10-10 06:10:49 +00:00
}
}
};
2011-10-10 10:05:39 +00:00
template <typename Impl, typename Name>
class FunctionStringNumNumToString : public IFunction
{
public:
/// Получить имя функции.
String getName() const
{
return Name::get();
}
/// Получить тип результата по типам аргументов. Если функция неприменима для данных аргументов - кинуть исключение.
DataTypePtr getReturnType(const DataTypes & arguments) const
{
if (arguments.size() != 3)
throw Exception("Number of arguments for function " + getName() + " doesn't match: passed "
2013-06-21 20:34:19 +00:00
+ toString(arguments.size()) + ", should be 3.",
2011-10-10 10:05:39 +00:00
ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH);
if (!typeid_cast<const DataTypeString *>(&*arguments[0]) && !typeid_cast<const DataTypeFixedString *>(&*arguments[0]))
2011-10-10 10:05:39 +00:00
throw Exception("Illegal type " + arguments[0]->getName() + " of argument of function " + getName(),
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
if (!arguments[1]->isNumeric() || !arguments[2]->isNumeric())
throw Exception("Illegal type " + (arguments[1]->isNumeric() ? arguments[2]->getName() : arguments[1]->getName()) + " of argument of function " + getName(),
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
return new DataTypeString;
}
/// Выполнить функцию над блоком.
void execute(Block & block, const ColumnNumbers & arguments, size_t result)
{
const ColumnPtr column_string = block.getByPosition(arguments[0]).column;
const ColumnPtr column_start = block.getByPosition(arguments[1]).column;
const ColumnPtr column_length = block.getByPosition(arguments[2]).column;
if (!column_start->isConst() || !column_length->isConst())
throw Exception("2nd and 3rd arguments of function " + getName() + " must be constants.");
FieldVisitorConvertToNumber<UInt64> converter;
Field start_field = (*block.getByPosition(arguments[1]).column)[0];
Field length_field = (*block.getByPosition(arguments[2]).column)[0];
UInt64 start = apply_visitor(converter, start_field);
UInt64 length = apply_visitor(converter, length_field);
2011-10-10 10:05:39 +00:00
if (start == 0)
throw Exception("Second argument of function substring must be greater than 0.", ErrorCodes::ARGUMENT_OUT_OF_BOUND);
if (const ColumnString * col = typeid_cast<const ColumnString *>(&*column_string))
2011-10-10 10:05:39 +00:00
{
ColumnString * col_res = new ColumnString;
block.getByPosition(result).column = col_res;
Impl::vector(col->getChars(), col->getOffsets(),
2011-10-10 10:05:39 +00:00
start, length,
col_res->getChars(), col_res->getOffsets());
2011-10-10 10:05:39 +00:00
}
else if (const ColumnFixedString * col = typeid_cast<const ColumnFixedString *>(&*column_string))
2011-10-10 10:05:39 +00:00
{
ColumnString * col_res = new ColumnString;
2011-10-10 10:05:39 +00:00
block.getByPosition(result).column = col_res;
Impl::vector_fixed(col->getChars(), col->getN(),
2011-10-10 10:05:39 +00:00
start, length,
col_res->getChars(), col_res->getOffsets());
2011-10-10 10:05:39 +00:00
}
else if (const ColumnConstString * col = typeid_cast<const ColumnConstString *>(&*column_string))
2011-10-10 10:05:39 +00:00
{
String res;
Impl::constant(col->getData(), start, length, res);
ColumnConstString * col_res = new ColumnConstString(col->size(), res);
block.getByPosition(result).column = col_res;
}
else
throw Exception("Illegal column " + block.getByPosition(arguments[0]).column->getName()
+ " of first argument of function " + getName(),
ErrorCodes::ILLEGAL_COLUMN);
}
};
struct NameEmpty { static const char * get() { return "empty"; } };
struct NameNotEmpty { static const char * get() { return "notEmpty"; } };
2011-10-03 04:06:34 +00:00
struct NameLength { static const char * get() { return "length"; } };
struct NameLengthUTF8 { static const char * get() { return "lengthUTF8"; } };
2011-10-03 05:29:11 +00:00
struct NameLower { static const char * get() { return "lower"; } };
struct NameUpper { static const char * get() { return "upper"; } };
struct NameLowerUTF8 { static const char * get() { return "lowerUTF8"; } };
struct NameUpperUTF8 { static const char * get() { return "upperUTF8"; } };
2011-10-10 06:10:49 +00:00
struct NameReverse { static const char * get() { return "reverse"; } };
struct NameReverseUTF8 { static const char * get() { return "reverseUTF8"; } };
struct NameConcat { static const char * get() { return "concat"; } };
2011-10-10 10:05:39 +00:00
struct NameSubstring { static const char * get() { return "substring"; } };
2011-10-10 12:26:26 +00:00
struct NameSubstringUTF8 { static const char * get() { return "substringUTF8"; } };
2011-10-03 04:06:34 +00:00
typedef FunctionStringOrArrayToT<EmptyImpl<false>, NameEmpty, UInt8> FunctionEmpty;
typedef FunctionStringOrArrayToT<EmptyImpl<true>, NameNotEmpty, UInt8> FunctionNotEmpty;
typedef FunctionStringOrArrayToT<LengthImpl, NameLength, UInt64> FunctionLength;
typedef FunctionStringOrArrayToT<LengthUTF8Impl, NameLengthUTF8, UInt64> FunctionLengthUTF8;
2011-10-03 05:29:11 +00:00
typedef FunctionStringToString<LowerUpperImpl<tolower>, NameLower> FunctionLower;
typedef FunctionStringToString<LowerUpperImpl<toupper>, NameUpper> FunctionUpper;
typedef FunctionStringToString<LowerUpperUTF8Impl<Poco::Unicode::toLower>, NameLowerUTF8> FunctionLowerUTF8;
typedef FunctionStringToString<LowerUpperUTF8Impl<Poco::Unicode::toUpper>, NameUpperUTF8> FunctionUpperUTF8;
2011-10-10 06:10:49 +00:00
typedef FunctionStringToString<ReverseImpl, NameReverse> FunctionReverse;
typedef FunctionStringToString<ReverseUTF8Impl, NameReverseUTF8> FunctionReverseUTF8;
typedef FunctionStringStringToString<ConcatImpl, NameConcat> FunctionConcat;
2011-10-10 10:05:39 +00:00
typedef FunctionStringNumNumToString<SubstringImpl, NameSubstring> FunctionSubstring;
2011-10-10 12:26:26 +00:00
typedef FunctionStringNumNumToString<SubstringUTF8Impl, NameSubstringUTF8> FunctionSubstringUTF8;
2011-10-03 04:06:34 +00:00
}