mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-12-11 08:52:06 +00:00
239 lines
7.6 KiB
C++
239 lines
7.6 KiB
C++
#include <Columns/ColumnString.h>
|
|
#include <Functions/FunctionFactory.h>
|
|
#include <Functions/FunctionStringToString.h>
|
|
#include <Common/StringUtils/StringUtils.h>
|
|
#include <Common/hex.h>
|
|
#include <base/find_symbols.h>
|
|
|
|
|
|
namespace DB
|
|
{
|
|
namespace ErrorCodes
|
|
{
|
|
extern const int ILLEGAL_TYPE_OF_ARGUMENT;
|
|
}
|
|
|
|
namespace
|
|
{
|
|
struct DecodeXMLComponentName
|
|
{
|
|
static constexpr auto name = "decodeXMLComponent";
|
|
};
|
|
|
|
class FunctionDecodeXMLComponentImpl
|
|
{
|
|
public:
|
|
static void vector(
|
|
const ColumnString::Chars & data,
|
|
const ColumnString::Offsets & offsets,
|
|
ColumnString::Chars & res_data,
|
|
ColumnString::Offsets & res_offsets)
|
|
{
|
|
/// The size of result is always not more than the size of source.
|
|
/// Because entities decodes to the shorter byte sequence.
|
|
/// Example: &#xx... &#xx... will decode to UTF-8 byte sequence not longer than 4 bytes.
|
|
res_data.resize(data.size());
|
|
|
|
size_t size = offsets.size();
|
|
res_offsets.resize(size);
|
|
|
|
size_t prev_offset = 0;
|
|
size_t res_offset = 0;
|
|
|
|
for (size_t i = 0; i < size; ++i)
|
|
{
|
|
const char * src_data = reinterpret_cast<const char *>(&data[prev_offset]);
|
|
size_t src_size = offsets[i] - prev_offset;
|
|
size_t dst_size = execute(src_data, src_size, reinterpret_cast<char *>(res_data.data() + res_offset));
|
|
|
|
res_offset += dst_size;
|
|
res_offsets[i] = res_offset;
|
|
prev_offset = offsets[i];
|
|
}
|
|
|
|
res_data.resize(res_offset);
|
|
}
|
|
|
|
[[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &)
|
|
{
|
|
throw Exception("Function decodeXMLComponent cannot work with FixedString argument", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
|
|
}
|
|
|
|
private:
|
|
static const int max_legal_unicode_value = 0x10FFFF;
|
|
static const int max_decimal_length_of_unicode_point = 7; /// 1114111
|
|
|
|
static size_t execute(const char * src, size_t src_size, char * dst)
|
|
{
|
|
const char * src_pos = src;
|
|
const char * src_end = src + src_size;
|
|
char * dst_pos = dst;
|
|
|
|
while (true)
|
|
{
|
|
const char * entity_pos = find_first_symbols<'&'>(src_pos, src_end);
|
|
|
|
if (entity_pos + strlen("lt;") >= src_end)
|
|
break;
|
|
|
|
/// Copy text between entities.
|
|
size_t bytes_to_copy = entity_pos - src_pos;
|
|
memcpySmallAllowReadWriteOverflow15(dst_pos, src_pos, bytes_to_copy);
|
|
dst_pos += bytes_to_copy;
|
|
src_pos = entity_pos;
|
|
|
|
++entity_pos;
|
|
|
|
const char * entity_end = find_first_symbols<';'>(entity_pos, src_end);
|
|
if (entity_end == src_end)
|
|
break;
|
|
|
|
bool parsed = false;
|
|
|
|
/// &#NNNN; or &#xNNNN;
|
|
uint32_t code_point = 0;
|
|
if (isValidNumericEntity(entity_pos, entity_end, code_point))
|
|
{
|
|
codePointToUTF8(code_point, dst_pos);
|
|
parsed = true;
|
|
}
|
|
else if (entity_end - entity_pos == 2)
|
|
{
|
|
if (memcmp(entity_pos, "lt", 2) == 0)
|
|
{
|
|
*dst_pos = '<';
|
|
++dst_pos;
|
|
parsed = true;
|
|
}
|
|
else if (memcmp(entity_pos, "gt", 2) == 0)
|
|
{
|
|
*dst_pos = '>';
|
|
++dst_pos;
|
|
parsed = true;
|
|
}
|
|
}
|
|
else if (entity_end - entity_pos == 3)
|
|
{
|
|
if (memcmp(entity_pos, "amp", 3) == 0)
|
|
{
|
|
*dst_pos = '&';
|
|
++dst_pos;
|
|
parsed = true;
|
|
}
|
|
}
|
|
else if (entity_end - entity_pos == 4)
|
|
{
|
|
if (memcmp(entity_pos, "quot", 4) == 0)
|
|
{
|
|
*dst_pos = '"';
|
|
++dst_pos;
|
|
parsed = true;
|
|
}
|
|
else if (memcmp(entity_pos, "apos", 4) == 0)
|
|
{
|
|
*dst_pos = '\'';
|
|
++dst_pos;
|
|
parsed = true;
|
|
}
|
|
}
|
|
|
|
if (parsed)
|
|
{
|
|
/// Skip the parsed entity.
|
|
src_pos = entity_end + 1;
|
|
}
|
|
else
|
|
{
|
|
/// Copy one byte as is and skip it.
|
|
*dst_pos = *src_pos;
|
|
++dst_pos;
|
|
++src_pos;
|
|
}
|
|
}
|
|
|
|
/// Copy the rest of the string.
|
|
if (src_pos < src_end)
|
|
{
|
|
size_t bytes_to_copy = src_end - src_pos;
|
|
memcpySmallAllowReadWriteOverflow15(dst_pos, src_pos, bytes_to_copy);
|
|
dst_pos += bytes_to_copy;
|
|
}
|
|
|
|
return dst_pos - dst;
|
|
}
|
|
|
|
static void codePointToUTF8(uint32_t code_point, char *& dst_pos)
|
|
{
|
|
if (code_point < (1 << 7))
|
|
{
|
|
dst_pos[0] = (code_point & 0x7F);
|
|
++dst_pos;
|
|
}
|
|
else if (code_point < (1 << 11))
|
|
{
|
|
dst_pos[0] = ((code_point >> 6) & 0x1F) + 0xC0;
|
|
dst_pos[1] = (code_point & 0x3F) + 0x80;
|
|
dst_pos += 2;
|
|
}
|
|
else if (code_point < (1 << 16))
|
|
{
|
|
dst_pos[0] = ((code_point >> 12) & 0x0F) + 0xE0;
|
|
dst_pos[1] = ((code_point >> 6) & 0x3F) + 0x80;
|
|
dst_pos[2] = (code_point & 0x3F) + 0x80;
|
|
dst_pos += 3;
|
|
}
|
|
else
|
|
{
|
|
dst_pos[0] = ((code_point >> 18) & 0x07) + 0xF0;
|
|
dst_pos[1] = ((code_point >> 12) & 0x3F) + 0x80;
|
|
dst_pos[2] = ((code_point >> 6) & 0x3F) + 0x80;
|
|
dst_pos[3] = (code_point & 0x3F) + 0x80;
|
|
dst_pos += 4;
|
|
}
|
|
}
|
|
|
|
static bool isValidNumericEntity(const char * src, const char * end, uint32_t & code_point)
|
|
{
|
|
if (src + strlen("#") >= end)
|
|
return false;
|
|
|
|
if (src[0] != '#' || (end - src > 1 + max_decimal_length_of_unicode_point))
|
|
return false;
|
|
|
|
if (src + 2 < end && (src[1] == 'x' || src[1] == 'X'))
|
|
{
|
|
src += 2;
|
|
for (; src < end; ++src)
|
|
{
|
|
if (!isHexDigit(*src))
|
|
return false;
|
|
code_point *= 16;
|
|
code_point += unhex(*src);
|
|
}
|
|
}
|
|
else
|
|
{
|
|
src += 1;
|
|
for (; src < end; ++src)
|
|
{
|
|
if (!isNumericASCII(*src))
|
|
return false;
|
|
code_point *= 10;
|
|
code_point += *src - '0';
|
|
}
|
|
}
|
|
|
|
return code_point <= max_legal_unicode_value;
|
|
}
|
|
};
|
|
|
|
using FunctionDecodeXMLComponent = FunctionStringToString<FunctionDecodeXMLComponentImpl, DecodeXMLComponentName>;
|
|
|
|
}
|
|
|
|
void registerFunctionDecodeXMLComponent(FunctionFactory & factory)
|
|
{
|
|
factory.registerFunction<FunctionDecodeXMLComponent>();
|
|
}
|
|
}
|