ClickHouse/src/Functions/FunctionsMultiStringFuzzySearch.h

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

143 lines
6.6 KiB
C++
Raw Normal View History

2019-03-29 01:02:05 +00:00
#pragma once
#include <Columns/ColumnArray.h>
#include <Columns/ColumnConst.h>
#include <Columns/ColumnString.h>
#include <Columns/ColumnVector.h>
#include <Columns/ColumnsNumber.h>
#include <DataTypes/DataTypeArray.h>
#include <DataTypes/DataTypeString.h>
#include <DataTypes/DataTypesNumber.h>
#include <Functions/FunctionHelpers.h>
2021-05-17 07:30:42 +00:00
#include <Functions/IFunction.h>
2019-03-29 01:02:05 +00:00
#include <IO/WriteHelpers.h>
#include <Interpreters/Context.h>
#include <optional>
namespace DB
{
namespace ErrorCodes
{
extern const int ILLEGAL_TYPE_OF_ARGUMENT;
extern const int ILLEGAL_COLUMN;
}
2022-07-14 11:00:51 +00:00
/**
* multiFuzzyMatchAny(haystack, [pattern_1, pattern_2, ..., pattern_n])
* multiFuzzyMatchAnyIndex(haystack, [pattern_1, pattern_2, ..., pattern_n])
* multiFuzzyMatchAllIndices(haystack, [pattern_1, pattern_2, ..., pattern_n])
*
*/
2019-03-29 01:02:05 +00:00
template <typename Impl>
class FunctionsMultiStringFuzzySearch : public IFunction
2019-03-29 01:02:05 +00:00
{
public:
static constexpr auto name = Impl::name;
2021-06-01 12:20:52 +00:00
static FunctionPtr create(ContextPtr context)
2019-03-29 01:02:05 +00:00
{
2022-06-24 13:20:03 +00:00
const auto & settings = context->getSettingsRef();
2023-02-08 13:07:27 +00:00
return std::make_shared<FunctionsMultiStringFuzzySearch>(settings.allow_hyperscan, settings.max_hyperscan_regexp_length, settings.max_hyperscan_regexp_total_length, settings.reject_expensive_hyperscan_regexps);
}
2023-02-08 13:07:27 +00:00
FunctionsMultiStringFuzzySearch(bool allow_hyperscan_, size_t max_hyperscan_regexp_length_, size_t max_hyperscan_regexp_total_length_, bool reject_expensive_hyperscan_regexps_)
: allow_hyperscan(allow_hyperscan_)
, max_hyperscan_regexp_length(max_hyperscan_regexp_length_)
, max_hyperscan_regexp_total_length(max_hyperscan_regexp_total_length_)
2023-02-08 13:07:27 +00:00
, reject_expensive_hyperscan_regexps(reject_expensive_hyperscan_regexps_)
2022-06-24 13:33:34 +00:00
{}
2019-03-29 01:02:05 +00:00
String getName() const override { return name; }
size_t getNumberOfArguments() const override { return 3; }
bool useDefaultImplementationForConstants() const override { return true; }
2021-06-22 16:21:23 +00:00
bool isSuitableForShortCircuitArgumentsExecution(const DataTypesWithConstInfo & /*arguments*/) const override { return true; }
ColumnNumbers getArgumentsThatAreAlwaysConstant() const override { return {1}; }
2019-03-29 01:02:05 +00:00
DataTypePtr getReturnTypeImpl(const DataTypes & arguments) const override
{
if (!isString(arguments[0]))
2022-06-24 13:33:34 +00:00
throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal type {} of argument of function {}", arguments[0]->getName(), getName());
2019-03-29 01:02:05 +00:00
if (!isUnsignedInteger(arguments[1]))
2022-06-24 13:33:34 +00:00
throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal type {} of argument of function {}", arguments[1]->getName(), getName());
2019-03-29 01:02:05 +00:00
const DataTypeArray * array_type = checkAndGetDataType<DataTypeArray>(arguments[2].get());
if (!array_type || !checkAndGetDataType<DataTypeString>(array_type->getNestedType().get()))
2022-06-24 13:33:34 +00:00
throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal type {} of argument of function {}", arguments[2]->getName(), getName());
return Impl::getReturnType();
2019-03-29 01:02:05 +00:00
}
ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t /*input_rows_count*/) const override
2019-03-29 01:02:05 +00:00
{
2022-07-06 14:29:42 +00:00
const ColumnPtr & haystack_ptr = arguments[0].column;
const ColumnPtr & edit_distance_ptr = arguments[1].column;
const ColumnPtr & needles_ptr = arguments[2].column;
2019-03-29 01:02:05 +00:00
2022-07-06 14:29:42 +00:00
const ColumnString * col_haystack_vector = checkAndGetColumn<ColumnString>(&*haystack_ptr);
const ColumnConst * col_haystack_const = checkAndGetColumnConst<ColumnString>(&*haystack_ptr);
assert(static_cast<bool>(col_haystack_vector) ^ static_cast<bool>(col_haystack_const));
2019-03-29 01:02:05 +00:00
UInt32 edit_distance = 0;
2022-07-06 14:29:42 +00:00
if (const auto * col_const_uint8 = checkAndGetColumnConst<ColumnUInt8>(edit_distance_ptr.get()))
2022-06-25 15:41:24 +00:00
edit_distance = col_const_uint8->getValue<UInt8>();
2022-07-06 14:29:42 +00:00
else if (const auto * col_const_uint16 = checkAndGetColumnConst<ColumnUInt16>(edit_distance_ptr.get()))
2022-06-25 15:41:24 +00:00
edit_distance = col_const_uint16->getValue<UInt16>();
2022-07-06 14:29:42 +00:00
else if (const auto * col_const_uint32 = checkAndGetColumnConst<ColumnUInt32>(edit_distance_ptr.get()))
2022-06-25 15:41:24 +00:00
edit_distance = col_const_uint32->getValue<UInt32>();
2019-03-29 01:02:05 +00:00
else
throw Exception(ErrorCodes::ILLEGAL_COLUMN,
"Illegal column {}. The number is not const or does not fit in UInt32",
arguments[1].column->getName());
2019-03-29 01:02:05 +00:00
const ColumnArray * col_needles_vector = checkAndGetColumn<ColumnArray>(needles_ptr.get());
2022-07-06 14:29:42 +00:00
const ColumnConst * col_needles_const = checkAndGetColumnConst<ColumnArray>(needles_ptr.get());
assert(static_cast<bool>(col_needles_vector) ^ static_cast<bool>(col_needles_const));
if (col_haystack_const && col_needles_vector)
throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Function '{}' doesn't support search with non-constant needles in constant haystack", name);
2019-03-29 01:02:05 +00:00
using ResultType = typename Impl::ResultType;
2019-03-29 01:02:05 +00:00
auto col_res = ColumnVector<ResultType>::create();
auto col_offsets = ColumnArray::ColumnOffsets::create();
2019-03-29 01:02:05 +00:00
auto & vec_res = col_res->getData();
auto & offsets_res = col_offsets->getData();
/// the implementations are responsible for resizing the output column
if (col_needles_const)
Impl::vectorConstant(
2022-07-06 21:36:14 +00:00
col_haystack_vector->getChars(), col_haystack_vector->getOffsets(),
col_needles_const->getValue<Array>(),
vec_res, offsets_res,
edit_distance,
2023-02-08 13:07:27 +00:00
allow_hyperscan, max_hyperscan_regexp_length, max_hyperscan_regexp_total_length, reject_expensive_hyperscan_regexps);
else
Impl::vectorVector(
2022-07-06 21:36:14 +00:00
col_haystack_vector->getChars(), col_haystack_vector->getOffsets(),
col_needles_vector->getData(), col_needles_vector->getOffsets(),
vec_res, offsets_res,
edit_distance,
2023-02-08 13:07:27 +00:00
allow_hyperscan, max_hyperscan_regexp_length, max_hyperscan_regexp_total_length, reject_expensive_hyperscan_regexps);
2019-03-29 01:02:05 +00:00
// the combination of const haystack + const needle is not implemented because
// useDefaultImplementationForConstants() == true makes upper layers convert both to
// non-const columns
if constexpr (Impl::is_column_array)
2020-10-18 19:00:13 +00:00
return ColumnArray::create(std::move(col_res), std::move(col_offsets));
else
2020-10-18 19:00:13 +00:00
return col_res;
2019-03-29 01:02:05 +00:00
}
private:
const bool allow_hyperscan;
2022-06-24 13:21:32 +00:00
const size_t max_hyperscan_regexp_length;
const size_t max_hyperscan_regexp_total_length;
2023-02-08 13:07:27 +00:00
const bool reject_expensive_hyperscan_regexps;
2019-03-29 01:02:05 +00:00
};
}