ClickHouse/dbms/Columns/ColumnDecimal.cpp

270 lines
7.3 KiB
C++
Raw Normal View History

#include <Common/Exception.h>
#include <Common/Arena.h>
#include <Common/SipHash.h>
#include <Common/assert_cast.h>
#include <Common/WeakHash.h>
2020-03-18 16:40:23 +00:00
#include <Common/HashTable/Hash.h>
2018-12-14 17:50:10 +00:00
#include <common/unaligned.h>
#include <IO/WriteHelpers.h>
#include <Columns/ColumnsCommon.h>
#include <Columns/ColumnDecimal.h>
#include <DataStreams/ColumnGathererStream.h>
2018-12-14 17:50:10 +00:00
template <typename T> bool decimalLess(T x, T y, UInt32 x_scale, UInt32 y_scale);
namespace DB
{
namespace ErrorCodes
{
extern const int PARAMETER_OUT_OF_BOUND;
extern const int SIZES_OF_COLUMNS_DOESNT_MATCH;
extern const int NOT_IMPLEMENTED;
2020-03-19 17:35:08 +00:00
extern const int LOGICAL_ERROR;
}
template <typename T>
2018-11-24 01:48:06 +00:00
int ColumnDecimal<T>::compareAt(size_t n, size_t m, const IColumn & rhs_, int) const
{
2019-08-07 13:53:46 +00:00
auto & other = static_cast<const Self &>(rhs_);
const T & a = data[n];
2018-08-29 18:47:19 +00:00
const T & b = other.data[m];
2019-08-07 14:12:57 +00:00
if (scale == other.scale)
return a > b ? 1 : (a < b ? -1 : 0);
return decimalLess<T>(b, a, other.scale, scale) ? 1 : (decimalLess<T>(a, b, scale, other.scale) ? -1 : 0);
}
template <typename T>
StringRef ColumnDecimal<T>::serializeValueIntoArena(size_t n, Arena & arena, char const *& begin) const
{
auto pos = arena.allocContinue(sizeof(T), begin);
memcpy(pos, &data[n], sizeof(T));
return StringRef(pos, sizeof(T));
}
template <typename T>
const char * ColumnDecimal<T>::deserializeAndInsertFromArena(const char * pos)
{
2018-12-14 17:50:10 +00:00
data.push_back(unalignedLoad<T>(pos));
return pos + sizeof(T);
}
template <typename T>
UInt64 ColumnDecimal<T>::get64(size_t n) const
{
if constexpr (sizeof(T) > sizeof(UInt64))
throw Exception(String("Method get64 is not supported for ") + getFamilyName(), ErrorCodes::NOT_IMPLEMENTED);
return static_cast<typename T::NativeType>(data[n]);
}
template <typename T>
void ColumnDecimal<T>::updateHashWithValue(size_t n, SipHash & hash) const
{
hash.update(data[n]);
}
template <typename T>
void ColumnDecimal<T>::updateWeakHash32(WeakHash32 & hash) const
{
auto s = data.size();
if (hash.getData().size() != s)
throw Exception("Size of WeakHash32 does not match size of column: column size is " + std::to_string(s) +
", hash size is " + std::to_string(hash.getData().size()), ErrorCodes::LOGICAL_ERROR);
2020-03-25 11:14:11 +00:00
const T * begin = data.data();
const T * end = begin + s;
2020-03-25 11:14:11 +00:00
UInt32 * hash_data = hash.getData().data();
while (begin < end)
{
2020-03-23 19:18:35 +00:00
*hash_data = intHashCRC32(*begin, *hash_data);
++begin;
++hash_data;
}
}
template <typename T>
void ColumnDecimal<T>::getPermutation(bool reverse, size_t limit, int , IColumn::Permutation & res) const
{
2018-08-29 18:47:19 +00:00
#if 1 /// TODO: perf test
if (data.size() <= std::numeric_limits<UInt32>::max())
{
2018-08-29 18:47:19 +00:00
PaddedPODArray<UInt32> tmp_res;
permutation(reverse, limit, tmp_res);
res.resize(tmp_res.size());
for (size_t i = 0; i < tmp_res.size(); ++i)
res[i] = tmp_res[i];
return;
}
2018-08-29 18:47:19 +00:00
#endif
permutation(reverse, limit, res);
}
template <typename T>
ColumnPtr ColumnDecimal<T>::permute(const IColumn::Permutation & perm, size_t limit) const
2018-08-29 18:47:19 +00:00
{
size_t size = limit ? std::min(data.size(), limit) : data.size();
if (perm.size() < size)
throw Exception("Size of permutation is less than required.", ErrorCodes::SIZES_OF_COLUMNS_DOESNT_MATCH);
auto res = this->create(size, scale);
typename Self::Container & res_data = res->getData();
for (size_t i = 0; i < size; ++i)
res_data[i] = data[perm[i]];
2019-05-31 08:55:01 +00:00
return res;
}
template <typename T>
MutableColumnPtr ColumnDecimal<T>::cloneResized(size_t size) const
{
auto res = this->create(0, scale);
if (size > 0)
{
auto & new_col = static_cast<Self &>(*res);
new_col.data.resize(size);
size_t count = std::min(this->size(), size);
memcpy(new_col.data.data(), data.data(), count * sizeof(data[0]));
if (size > count)
2018-08-29 18:47:19 +00:00
{
void * tail = &new_col.data[count];
memset(tail, 0, (size - count) * sizeof(T));
}
}
2019-05-31 08:55:01 +00:00
return res;
}
template <typename T>
void ColumnDecimal<T>::insertData(const char * src, size_t /*length*/)
{
T tmp;
memcpy(&tmp, src, sizeof(T));
data.emplace_back(tmp);
}
template <typename T>
void ColumnDecimal<T>::insertRangeFrom(const IColumn & src, size_t start, size_t length)
{
const ColumnDecimal & src_vec = assert_cast<const ColumnDecimal &>(src);
if (start + length > src_vec.data.size())
2018-08-29 18:47:19 +00:00
throw Exception("Parameters start = " + toString(start) + ", length = " + toString(length) +
" are out of bound in ColumnDecimal<T>::insertRangeFrom method (data.size() = " + toString(src_vec.data.size()) + ").",
ErrorCodes::PARAMETER_OUT_OF_BOUND);
size_t old_size = data.size();
data.resize(old_size + length);
2019-03-07 20:04:59 +00:00
memcpy(data.data() + old_size, &src_vec.data[start], length * sizeof(data[0]));
}
template <typename T>
ColumnPtr ColumnDecimal<T>::filter(const IColumn::Filter & filt, ssize_t result_size_hint) const
{
size_t size = data.size();
if (size != filt.size())
throw Exception("Size of filter doesn't match size of column.", ErrorCodes::SIZES_OF_COLUMNS_DOESNT_MATCH);
auto res = this->create(0, scale);
Container & res_data = res->getData();
if (result_size_hint)
res_data.reserve(result_size_hint > 0 ? result_size_hint : size);
const UInt8 * filt_pos = filt.data();
const UInt8 * filt_end = filt_pos + size;
const T * data_pos = data.data();
while (filt_pos < filt_end)
{
if (*filt_pos)
res_data.push_back(*data_pos);
++filt_pos;
++data_pos;
}
2019-05-31 08:55:01 +00:00
return res;
}
template <typename T>
ColumnPtr ColumnDecimal<T>::index(const IColumn & indexes, size_t limit) const
{
return selectIndexImpl(*this, indexes, limit);
}
template <typename T>
ColumnPtr ColumnDecimal<T>::replicate(const IColumn::Offsets & offsets) const
{
size_t size = data.size();
if (size != offsets.size())
throw Exception("Size of offsets doesn't match size of column.", ErrorCodes::SIZES_OF_COLUMNS_DOESNT_MATCH);
auto res = this->create(0, scale);
if (0 == size)
return res;
typename Self::Container & res_data = res->getData();
res_data.reserve(offsets.back());
IColumn::Offset prev_offset = 0;
for (size_t i = 0; i < size; ++i)
{
size_t size_to_replicate = offsets[i] - prev_offset;
prev_offset = offsets[i];
for (size_t j = 0; j < size_to_replicate; ++j)
res_data.push_back(data[i]);
}
2019-05-31 08:55:01 +00:00
return res;
}
template <typename T>
void ColumnDecimal<T>::gather(ColumnGathererStream & gatherer)
{
gatherer.gather(*this);
}
template <typename T>
void ColumnDecimal<T>::getExtremes(Field & min, Field & max) const
{
2020-03-08 21:40:00 +00:00
if (data.empty())
{
2018-11-20 20:09:20 +00:00
min = NearestFieldType<T>(0, scale);
max = NearestFieldType<T>(0, scale);
return;
}
T cur_min = data[0];
T cur_max = data[0];
for (const T & x : data)
{
if (x < cur_min)
cur_min = x;
else if (x > cur_max)
cur_max = x;
}
2018-11-20 20:09:20 +00:00
min = NearestFieldType<T>(cur_min, scale);
max = NearestFieldType<T>(cur_max, scale);
}
template class ColumnDecimal<Decimal32>;
template class ColumnDecimal<Decimal64>;
template class ColumnDecimal<Decimal128>;
}