ClickHouse/src/Functions/dateDiff.cpp

227 lines
9.8 KiB
C++
Raw Normal View History

#include <DataTypes/DataTypeDateTime.h>
#include <DataTypes/DataTypeDateTime64.h>
#include <DataTypes/DataTypesNumber.h>
#include <Columns/ColumnString.h>
#include <Columns/ColumnsNumber.h>
#include <Functions/IFunctionImpl.h>
#include <Functions/FunctionHelpers.h>
#include <Functions/FunctionFactory.h>
#include <Functions/extractTimeZoneFromFunctionArguments.h>
#include <Functions/DateTimeTransforms.h>
#include <IO/WriteHelpers.h>
#include <common/find_symbols.h>
#include <type_traits>
namespace DB
{
namespace ErrorCodes
{
extern const int NUMBER_OF_ARGUMENTS_DOESNT_MATCH;
extern const int ILLEGAL_TYPE_OF_ARGUMENT;
extern const int ILLEGAL_COLUMN;
extern const int BAD_ARGUMENTS;
}
2020-09-07 18:00:37 +00:00
namespace
{
/** dateDiff('unit', t1, t2, [timezone])
* t1 and t2 can be Date or DateTime
*
* If timezone is specified, it applied to both arguments.
* If not, timezones from datatypes t1 and t2 are used.
* If that timezones are not the same, the result is unspecified.
*
* Timezone matters because days can have different length.
*/
class FunctionDateDiff : public IFunction
{
public:
static constexpr auto name = "dateDiff";
static FunctionPtr create(const Context &) { return std::make_shared<FunctionDateDiff>(); }
String getName() const override
{
return name;
}
bool isVariadic() const override { return true; }
size_t getNumberOfArguments() const override { return 0; }
DataTypePtr getReturnTypeImpl(const DataTypes & arguments) const override
{
if (arguments.size() != 3 && arguments.size() != 4)
throw Exception("Number of arguments for function " + getName() + " doesn't match: passed "
+ toString(arguments.size()) + ", should be 3 or 4",
ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH);
if (!isString(arguments[0]))
throw Exception("First argument for function " + getName() + " (unit) must be String",
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
if (!isDateOrDateTime(arguments[1]))
throw Exception("Second argument for function " + getName() + " must be Date or DateTime",
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
if (!isDateOrDateTime(arguments[2]))
throw Exception("Third argument for function " + getName() + " must be Date or DateTime",
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
if (arguments.size() == 4 && !isString(arguments[3]))
throw Exception("Fourth argument for function " + getName() + " (timezone) must be String",
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
return std::make_shared<DataTypeInt64>();
}
bool useDefaultImplementationForConstants() const override { return true; }
ColumnNumbers getArgumentsThatAreAlwaysConstant() const override { return {0, 3}; }
ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t input_rows_count) const override
{
2020-10-17 21:41:50 +00:00
const auto * unit_column = checkAndGetColumnConst<ColumnString>(arguments[0].column.get());
if (!unit_column)
throw Exception("First argument for function " + getName() + " must be constant String", ErrorCodes::ILLEGAL_COLUMN);
String unit = Poco::toLower(unit_column->getValue<String>());
2020-10-17 21:41:50 +00:00
const IColumn & x = *arguments[1].column;
const IColumn & y = *arguments[2].column;
size_t rows = input_rows_count;
auto res = ColumnInt64::create(rows);
Extended range of DateTime64 to years 1925 - 2238 The Year 1925 is a starting point because most of the timezones switched to saner (mostly 15-minutes based) offsets somewhere during 1924 or before. And that significantly simplifies implementation. 2238 is to simplify arithmetics for sanitizing LUT index access; there are less than 0x1ffff days from 1925. * Extended DateLUTImpl internal LUT to 0x1ffff items, some of which represent negative (pre-1970) time values. As a collateral benefit, Date now correctly supports dates up to 2149 (instead of 2106). * Added a new strong typedef ExtendedDayNum, which represents dates pre-1970 and post 2149. * Functions that used to return DayNum now return ExtendedDayNum. * Refactored DateLUTImpl to untie DayNum from the dual role of being a value and an index (due to negative time). Index is now a different type LUTIndex with explicit conversion functions from DatNum, time_t, and ExtendedDayNum. * Updated DateLUTImpl to properly support values close to epoch start (1970-01-01 00:00), including negative ones. * Reduced resolution of DateLUTImpl::Values::time_at_offset_change to multiple of 15-minutes to allow storing 64-bits of time_t in DateLUTImpl::Value while keeping same size. * Minor performance updates to DateLUTImpl when building month LUT by skipping non-start-of-month days. * Fixed extractTimeZoneFromFunctionArguments to work correctly with DateTime64. * New unit-tests and stateless integration tests for both DateTime and DateTime64.
2020-04-17 13:26:44 +00:00
const auto & timezone_x = extractTimeZoneFromFunctionArguments(arguments, 3, 1);
const auto & timezone_y = extractTimeZoneFromFunctionArguments(arguments, 3, 2);
if (unit == "year" || unit == "yy" || unit == "yyyy")
dispatchForColumns<ToRelativeYearNumImpl>(x, y, timezone_x, timezone_y, res->getData());
else if (unit == "quarter" || unit == "qq" || unit == "q")
dispatchForColumns<ToRelativeQuarterNumImpl>(x, y, timezone_x, timezone_y, res->getData());
else if (unit == "month" || unit == "mm" || unit == "m")
dispatchForColumns<ToRelativeMonthNumImpl>(x, y, timezone_x, timezone_y, res->getData());
else if (unit == "week" || unit == "wk" || unit == "ww")
dispatchForColumns<ToRelativeWeekNumImpl>(x, y, timezone_x, timezone_y, res->getData());
else if (unit == "day" || unit == "dd" || unit == "d")
dispatchForColumns<ToRelativeDayNumImpl>(x, y, timezone_x, timezone_y, res->getData());
else if (unit == "hour" || unit == "hh")
dispatchForColumns<ToRelativeHourNumImpl>(x, y, timezone_x, timezone_y, res->getData());
else if (unit == "minute" || unit == "mi" || unit == "n")
dispatchForColumns<ToRelativeMinuteNumImpl>(x, y, timezone_x, timezone_y, res->getData());
else if (unit == "second" || unit == "ss" || unit == "s")
dispatchForColumns<ToRelativeSecondNumImpl>(x, y, timezone_x, timezone_y, res->getData());
else
throw Exception("Function " + getName() + " does not support '" + unit + "' unit", ErrorCodes::BAD_ARGUMENTS);
2020-10-17 21:41:50 +00:00
return res;
}
private:
template <typename Transform>
void dispatchForColumns(
const IColumn & x, const IColumn & y,
const DateLUTImpl & timezone_x, const DateLUTImpl & timezone_y,
2020-07-21 13:58:07 +00:00
ColumnInt64::Container & result) const
{
2020-04-22 08:31:10 +00:00
if (const auto * x_vec_16 = checkAndGetColumn<ColumnUInt16>(&x))
dispatchForSecondColumn<Transform>(*x_vec_16, y, timezone_x, timezone_y, result);
2020-04-22 08:31:10 +00:00
else if (const auto * x_vec_32 = checkAndGetColumn<ColumnUInt32>(&x))
dispatchForSecondColumn<Transform>(*x_vec_32, y, timezone_x, timezone_y, result);
2020-04-22 08:31:10 +00:00
else if (const auto * x_const_16 = checkAndGetColumnConst<ColumnUInt16>(&x))
dispatchConstForSecondColumn<Transform>(x_const_16->getValue<UInt16>(), y, timezone_x, timezone_y, result);
2020-04-22 08:31:10 +00:00
else if (const auto * x_const_32 = checkAndGetColumnConst<ColumnUInt32>(&x))
dispatchConstForSecondColumn<Transform>(x_const_32->getValue<UInt32>(), y, timezone_x, timezone_y, result);
else
throw Exception("Illegal column for first argument of function " + getName() + ", must be Date or DateTime", ErrorCodes::ILLEGAL_COLUMN);
}
template <typename Transform, typename T1>
void dispatchForSecondColumn(
const ColumnVector<T1> & x, const IColumn & y,
const DateLUTImpl & timezone_x, const DateLUTImpl & timezone_y,
2020-07-21 13:58:07 +00:00
ColumnInt64::Container & result) const
{
2020-04-22 08:31:10 +00:00
if (const auto * y_vec_16 = checkAndGetColumn<ColumnUInt16>(&y))
vectorVector<Transform>(x, *y_vec_16, timezone_x, timezone_y, result);
2020-04-22 08:31:10 +00:00
else if (const auto * y_vec_32 = checkAndGetColumn<ColumnUInt32>(&y))
vectorVector<Transform>(x, *y_vec_32, timezone_x, timezone_y, result);
2020-04-22 08:31:10 +00:00
else if (const auto * y_const_16 = checkAndGetColumnConst<ColumnUInt16>(&y))
vectorConstant<Transform>(x, y_const_16->getValue<UInt16>(), timezone_x, timezone_y, result);
2020-04-22 08:31:10 +00:00
else if (const auto * y_const_32 = checkAndGetColumnConst<ColumnUInt32>(&y))
vectorConstant<Transform>(x, y_const_32->getValue<UInt32>(), timezone_x, timezone_y, result);
else
throw Exception("Illegal column for second argument of function " + getName() + ", must be Date or DateTime", ErrorCodes::ILLEGAL_COLUMN);
}
template <typename Transform, typename T1>
void dispatchConstForSecondColumn(
T1 x, const IColumn & y,
const DateLUTImpl & timezone_x, const DateLUTImpl & timezone_y,
2020-07-21 13:58:07 +00:00
ColumnInt64::Container & result) const
{
2020-04-22 08:31:10 +00:00
if (const auto * y_vec_16 = checkAndGetColumn<ColumnUInt16>(&y))
constantVector<Transform>(x, *y_vec_16, timezone_x, timezone_y, result);
2020-04-22 08:31:10 +00:00
else if (const auto * y_vec_32 = checkAndGetColumn<ColumnUInt32>(&y))
constantVector<Transform>(x, *y_vec_32, timezone_x, timezone_y, result);
else
throw Exception("Illegal column for second argument of function " + getName() + ", must be Date or DateTime", ErrorCodes::ILLEGAL_COLUMN);
}
template <typename Transform, typename T1, typename T2>
void vectorVector(
const ColumnVector<T1> & x, const ColumnVector<T2> & y,
const DateLUTImpl & timezone_x, const DateLUTImpl & timezone_y,
2020-07-21 13:58:07 +00:00
ColumnInt64::Container & result) const
{
const auto & x_data = x.getData();
const auto & y_data = y.getData();
for (size_t i = 0, size = x.size(); i < size; ++i)
result[i] = calculate<Transform>(x_data[i], y_data[i], timezone_x, timezone_y);
}
template <typename Transform, typename T1, typename T2>
void vectorConstant(
const ColumnVector<T1> & x, T2 y,
const DateLUTImpl & timezone_x, const DateLUTImpl & timezone_y,
2020-07-21 13:58:07 +00:00
ColumnInt64::Container & result) const
{
const auto & x_data = x.getData();
for (size_t i = 0, size = x.size(); i < size; ++i)
result[i] = calculate<Transform>(x_data[i], y, timezone_x, timezone_y);
}
template <typename Transform, typename T1, typename T2>
void constantVector(
T1 x, const ColumnVector<T2> & y,
const DateLUTImpl & timezone_x, const DateLUTImpl & timezone_y,
2020-07-21 13:58:07 +00:00
ColumnInt64::Container & result) const
{
const auto & y_data = y.getData();
for (size_t i = 0, size = y.size(); i < size; ++i)
result[i] = calculate<Transform>(x, y_data[i], timezone_x, timezone_y);
}
template <typename Transform, typename T1, typename T2>
2020-07-21 13:58:07 +00:00
Int64 calculate(T1 x, T2 y, const DateLUTImpl & timezone_x, const DateLUTImpl & timezone_y) const
{
return Int64(Transform::execute(y, timezone_y))
- Int64(Transform::execute(x, timezone_x));
}
};
2020-09-07 18:00:37 +00:00
}
void registerFunctionDateDiff(FunctionFactory & factory)
{
factory.registerFunction<FunctionDateDiff>(FunctionFactory::CaseInsensitive);
}
}