ClickHouse/src/Functions/toFixedString.h

161 lines
6.1 KiB
C++
Raw Normal View History

2020-10-10 18:37:02 +00:00
#pragma once
2021-05-17 07:30:42 +00:00
#include <Functions/IFunction.h>
2020-06-21 18:57:20 +00:00
#include <DataTypes/DataTypeString.h>
#include <DataTypes/DataTypeFixedString.h>
#include <DataTypes/DataTypesNumber.h>
#include <Columns/ColumnString.h>
#include <Columns/ColumnFixedString.h>
2020-11-05 19:09:17 +00:00
#include <Columns/ColumnsNumber.h>
#include <Columns/ColumnNullable.h>
2020-06-21 18:57:20 +00:00
#include <IO/WriteHelpers.h>
2020-06-21 10:54:28 +00:00
namespace DB
{
2020-06-21 18:57:20 +00:00
namespace ErrorCodes
{
extern const int ILLEGAL_COLUMN;
extern const int TOO_LARGE_STRING_SIZE;
extern const int NOT_IMPLEMENTED;
}
2020-11-05 19:09:17 +00:00
enum class ConvertToFixedStringExceptionMode
{
Throw,
Null
};
2020-06-21 18:57:20 +00:00
2020-06-21 10:54:28 +00:00
/** Conversion to fixed string is implemented only for strings.
*/
class FunctionToFixedString : public IFunction
{
public:
static constexpr auto name = "toFixedString";
2021-06-01 12:20:52 +00:00
static FunctionPtr create(ContextPtr) { return std::make_shared<FunctionToFixedString>(); }
2020-06-21 10:54:28 +00:00
static FunctionPtr create() { return std::make_shared<FunctionToFixedString>(); }
String getName() const override
{
return name;
}
size_t getNumberOfArguments() const override { return 2; }
bool isInjective(const ColumnsWithTypeAndName &) const override { return true; }
2020-06-21 10:54:28 +00:00
2020-06-21 18:57:20 +00:00
DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override
{
if (!isUnsignedInteger(arguments[1].type))
throw Exception("Second argument for function " + getName() + " must be unsigned integer", ErrorCodes::ILLEGAL_COLUMN);
if (!arguments[1].column)
throw Exception("Second argument for function " + getName() + " must be constant", ErrorCodes::ILLEGAL_COLUMN);
if (!isStringOrFixedString(arguments[0].type))
throw Exception(getName() + " is only implemented for types String and FixedString", ErrorCodes::NOT_IMPLEMENTED);
const size_t n = arguments[1].column->getUInt(0);
return std::make_shared<DataTypeFixedString>(n);
}
2020-06-21 10:54:28 +00:00
bool useDefaultImplementationForConstants() const override { return true; }
ColumnNumbers getArgumentsThatAreAlwaysConstant() const override { return {1}; }
ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t /*input_rows_count*/) const override
2020-06-21 18:57:20 +00:00
{
2020-10-17 14:23:37 +00:00
const auto n = arguments[1].column->getUInt(0);
2020-11-05 19:09:17 +00:00
return executeForN<ConvertToFixedStringExceptionMode::Throw>(arguments, n);
2020-06-21 18:57:20 +00:00
}
2020-11-05 19:09:17 +00:00
template<ConvertToFixedStringExceptionMode exception_mode>
static ColumnPtr executeForN(const ColumnsWithTypeAndName & arguments, const size_t n)
2020-06-21 18:57:20 +00:00
{
2020-10-17 14:23:37 +00:00
const auto & column = arguments[0].column;
2020-06-21 18:57:20 +00:00
2020-11-05 19:09:17 +00:00
ColumnUInt8::MutablePtr col_null_map_to;
ColumnUInt8::Container * vec_null_map_to [[maybe_unused]] = nullptr;
if constexpr (exception_mode == ConvertToFixedStringExceptionMode::Null)
{
col_null_map_to = ColumnUInt8::create(column->size(), false);
vec_null_map_to = &col_null_map_to->getData();
}
2020-10-17 14:23:37 +00:00
if (const auto * column_string = checkAndGetColumn<ColumnString>(column.get()))
2020-06-21 18:57:20 +00:00
{
auto column_fixed = ColumnFixedString::create(n);
auto & out_chars = column_fixed->getChars();
const auto & in_chars = column_string->getChars();
const auto & in_offsets = column_string->getOffsets();
out_chars.resize_fill(in_offsets.size() * n);
for (size_t i = 0; i < in_offsets.size(); ++i)
{
const size_t off = i ? in_offsets[i - 1] : 0;
const size_t len = in_offsets[i] - off - 1;
if (len > n)
2020-11-05 19:09:17 +00:00
{
if constexpr (exception_mode == ConvertToFixedStringExceptionMode::Throw)
{
throw Exception("String too long for type FixedString(" + toString(n) + ")",
ErrorCodes::TOO_LARGE_STRING_SIZE);
}
else
{
(*vec_null_map_to)[i] = true;
continue;
}
}
2020-06-21 18:57:20 +00:00
memcpy(&out_chars[i * n], &in_chars[off], len);
}
2020-06-21 10:54:28 +00:00
2020-11-05 19:09:17 +00:00
if constexpr (exception_mode == ConvertToFixedStringExceptionMode::Null)
return ColumnNullable::create(std::move(column_fixed), std::move(col_null_map_to));
else
2020-11-30 08:26:00 +00:00
return column_fixed;
2020-06-21 18:57:20 +00:00
}
2020-10-17 14:23:37 +00:00
else if (const auto * column_fixed_string = checkAndGetColumn<ColumnFixedString>(column.get()))
2020-06-21 18:57:20 +00:00
{
const auto src_n = column_fixed_string->getN();
if (src_n > n)
2020-11-05 19:09:17 +00:00
{
if constexpr (exception_mode == ConvertToFixedStringExceptionMode::Throw)
{
throw Exception{"String too long for type FixedString(" + toString(n) + ")", ErrorCodes::TOO_LARGE_STRING_SIZE};
}
else
{
auto column_fixed = ColumnFixedString::create(n);
std::fill(vec_null_map_to->begin(), vec_null_map_to->end(), true);
return ColumnNullable::create(column_fixed->cloneResized(column->size()), std::move(col_null_map_to));
}
}
2020-06-21 18:57:20 +00:00
auto column_fixed = ColumnFixedString::create(n);
auto & out_chars = column_fixed->getChars();
const auto & in_chars = column_fixed_string->getChars();
const auto size = column_fixed_string->size();
out_chars.resize_fill(size * n);
for (size_t i = 0; i < size; ++i)
memcpy(&out_chars[i * n], &in_chars[i * src_n], src_n);
2020-10-17 14:23:37 +00:00
return column_fixed;
2020-06-21 18:57:20 +00:00
}
else
2020-11-05 19:09:17 +00:00
{
if constexpr (exception_mode == ConvertToFixedStringExceptionMode::Throw)
throw Exception("Unexpected column: " + column->getName(), ErrorCodes::ILLEGAL_COLUMN);
else
{
auto column_fixed = ColumnFixedString::create(n);
std::fill(vec_null_map_to->begin(), vec_null_map_to->end(), true);
return ColumnNullable::create(column_fixed->cloneResized(column->size()), std::move(col_null_map_to));
}
}
2020-06-21 18:57:20 +00:00
}
2020-06-21 10:54:28 +00:00
};
}