1
0
mirror of https://github.com/mariadb-corporation/mariadb-columnstore-engine.git synced 2025-08-01 06:46:55 +03:00

MCOL-1201 Add test UDAF back in after rebase

This commit is contained in:
David Hall
2018-05-11 09:53:49 -05:00
parent 6fa7dded6f
commit c67ac7699e
4 changed files with 725 additions and 0 deletions

257
utils/udfsdk/avgx.cpp Normal file
View File

@ -0,0 +1,257 @@
/* Copyright (C) 2017 MariaDB Corporaton
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; version 2 of
the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
MA 02110-1301, USA. */
#include <sstream>
#include <cstring>
#include <typeinfo>
#include "avgx.h"
#include "bytestream.h"
#include "objectreader.h"
using namespace mcsv1sdk;
#define DATATYPE double
// Use the simple data model
struct avgx_data
{
double sum;
uint64_t cnt;
};
mcsv1_UDAF::ReturnCode avgx::init(mcsv1Context* context,
ColumnDatum* colTypes)
{
if (context->getParameterCount() != 1)
{
// The error message will be prepended with
// "The storage engine for the table doesn't support "
context->setErrorMessage("avgx() with other than 1 arguments");
return mcsv1_UDAF::ERROR;
}
if (!(isNumeric(colTypes[0].dataType)))
{
// The error message will be prepended with
// "The storage engine for the table doesn't support "
context->setErrorMessage("avgx() with a non-numeric x argument");
return mcsv1_UDAF::ERROR;
}
context->setUserDataSize(sizeof(avgx_data));
context->setResultType(CalpontSystemCatalog::DOUBLE);
context->setColWidth(8);
context->setScale(colTypes[0].scale + 4);
context->setPrecision(19);
context->setRunFlag(mcsv1sdk::UDAF_IGNORE_NULLS);
return mcsv1_UDAF::SUCCESS;
}
mcsv1_UDAF::ReturnCode avgx::reset(mcsv1Context* context)
{
struct avgx_data* data = (struct avgx_data*)context->getUserData()->data;
data->sum = 0;
data->cnt = 0;
return mcsv1_UDAF::SUCCESS;
}
mcsv1_UDAF::ReturnCode avgx::nextValue(mcsv1Context* context, ColumnDatum* valsIn)
{
static_any::any& valIn_x = valsIn[0].columnData;
struct avgx_data* data = (struct avgx_data*)context->getUserData()->data;
DATATYPE val = 0.0;
if (valIn_x.empty())
{
return mcsv1_UDAF::SUCCESS; // Ought not happen when UDAF_IGNORE_NULLS is on.
}
if (valIn_x.compatible(longTypeId))
{
val = valIn_x.cast<long>();
}
else if (valIn_x.compatible(charTypeId))
{
val = valIn_x.cast<char>();
}
else if (valIn_x.compatible(scharTypeId))
{
val = valIn_x.cast<signed char>();
}
else if (valIn_x.compatible(shortTypeId))
{
val = valIn_x.cast<short>();
}
else if (valIn_x.compatible(intTypeId))
{
val = valIn_x.cast<int>();
}
else if (valIn_x.compatible(longTypeId))
{
val = valIn_x.cast<long>();
}
else if (valIn_x.compatible(llTypeId))
{
val = valIn_x.cast<long long>();
}
else if (valIn_x.compatible(ucharTypeId))
{
val = valIn_x.cast<unsigned char>();
}
else if (valIn_x.compatible(ushortTypeId))
{
val = valIn_x.cast<unsigned short>();
}
else if (valIn_x.compatible(uintTypeId))
{
val = valIn_x.cast<unsigned int>();
}
else if (valIn_x.compatible(ulongTypeId))
{
val = valIn_x.cast<unsigned long>();
}
else if (valIn_x.compatible(ullTypeId))
{
val = valIn_x.cast<unsigned long long>();
}
else if (valIn_x.compatible(floatTypeId))
{
val = valIn_x.cast<float>();
}
else if (valIn_x.compatible(doubleTypeId))
{
val = valIn_x.cast<double>();
}
// For decimal types, we need to move the decimal point.
uint32_t scale = valsIn[0].scale;
if (val != 0 && scale > 0)
{
val /= pow(10.0, (double)scale);
}
data->sum += val;
++data->cnt;
return mcsv1_UDAF::SUCCESS;
}
mcsv1_UDAF::ReturnCode avgx::subEvaluate(mcsv1Context* context, const UserData* userDataIn)
{
if (!userDataIn)
{
return mcsv1_UDAF::SUCCESS;
}
struct avgx_data* outData = (struct avgx_data*)context->getUserData()->data;
struct avgx_data* inData = (struct avgx_data*)userDataIn->data;
outData->sum += inData->sum;
outData->cnt += inData->cnt;
return mcsv1_UDAF::SUCCESS;
}
mcsv1_UDAF::ReturnCode avgx::evaluate(mcsv1Context* context, static_any::any& valOut)
{
struct avgx_data* data = (struct avgx_data*)context->getUserData()->data;
valOut = data->sum / (double)data->cnt;
return mcsv1_UDAF::SUCCESS;
}
mcsv1_UDAF::ReturnCode avgx::dropValue(mcsv1Context* context, ColumnDatum* valsDropped)
{
static_any::any& valIn_x = valsDropped[0].columnData;
struct avgx_data* data = (struct avgx_data*)context->getUserData()->data;
DATATYPE val = 0.0;
if (valIn_x.empty())
{
return mcsv1_UDAF::SUCCESS; // Ought not happen when UDAF_IGNORE_NULLS is on.
}
if (valIn_x.compatible(charTypeId))
{
val = valIn_x.cast<char>();
}
else if (valIn_x.compatible(scharTypeId))
{
val = valIn_x.cast<signed char>();
}
else if (valIn_x.compatible(shortTypeId))
{
val = valIn_x.cast<short>();
}
else if (valIn_x.compatible(intTypeId))
{
val = valIn_x.cast<int>();
}
else if (valIn_x.compatible(longTypeId))
{
val = valIn_x.cast<long>();
}
else if (valIn_x.compatible(llTypeId))
{
val = valIn_x.cast<long long>();
}
else if (valIn_x.compatible(ucharTypeId))
{
val = valIn_x.cast<unsigned char>();
}
else if (valIn_x.compatible(ushortTypeId))
{
val = valIn_x.cast<unsigned short>();
}
else if (valIn_x.compatible(uintTypeId))
{
val = valIn_x.cast<unsigned int>();
}
else if (valIn_x.compatible(ulongTypeId))
{
val = valIn_x.cast<unsigned long>();
}
else if (valIn_x.compatible(ullTypeId))
{
val = valIn_x.cast<unsigned long long>();
}
else if (valIn_x.compatible(floatTypeId))
{
val = valIn_x.cast<float>();
}
else if (valIn_x.compatible(doubleTypeId))
{
val = valIn_x.cast<double>();
}
// For decimal types, we need to move the decimal point.
uint32_t scale = valsDropped[0].scale;
if (val != 0 && scale > 0)
{
val /= pow(10.0, (double)scale);
}
data->sum -= val;
--data->cnt;
return mcsv1_UDAF::SUCCESS;
}

99
utils/udfsdk/avgx.h Normal file
View File

@ -0,0 +1,99 @@
/* Copyright (C) 2017 MariaDB Corporaton
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; version 2 of
the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
MA 02110-1301, USA. */
/***********************************************************************
* $Id$
*
* avgx.h
***********************************************************************/
/**
* Columnstore interface for for the avgx function
*
*
* CREATE AGGREGATE FUNCTION avgx returns REAL soname
* 'libudf_mysql.so';
*
*/
#ifndef HEADER_avgx
#define HEADER_avgx
#include <cstdlib>
#include <string>
#include <vector>
#include <boost/any.hpp>
#ifdef _MSC_VER
#include <unordered_map>
#else
#include <tr1/unordered_map>
#endif
#include "mcsv1_udaf.h"
#include "calpontsystemcatalog.h"
#include "windowfunctioncolumn.h"
using namespace execplan;
#if defined(_MSC_VER) && defined(xxxRGNODE_DLLEXPORT)
#define EXPORT __declspec(dllexport)
#else
#define EXPORT
#endif
namespace mcsv1sdk
{
// Override mcsv1_UDAF to build your User Defined Aggregate (UDAF) and/or
// User Defined Analytic Function (UDAnF).
// These will be singleton classes, so don't put any instance
// specific data in here. All instance data is stored in mcsv1Context
// passed to each user function and retrieved by the getUserData() method.
//
// Each API function returns a ReturnCode. If ERROR is returned at any time,
// the query is aborted, getInterrupted() will begin to return true and the
// message set in config->setErrorMessage() is returned to MariaDB.
// Return the avgx value of the dataset
class avgx : public mcsv1_UDAF
{
public:
// Defaults OK
avgx() : mcsv1_UDAF() {};
virtual ~avgx() {};
virtual ReturnCode init(mcsv1Context* context,
ColumnDatum* colTypes);
virtual ReturnCode reset(mcsv1Context* context);
virtual ReturnCode nextValue(mcsv1Context* context, ColumnDatum* valsIn);
virtual ReturnCode subEvaluate(mcsv1Context* context, const UserData* valIn);
virtual ReturnCode evaluate(mcsv1Context* context, static_any::any& valOut);
virtual ReturnCode dropValue(mcsv1Context* context, ColumnDatum* valsDropped);
protected:
};
}; // namespace
#undef EXPORT
#endif // HEADER_.h

270
utils/udfsdk/regr_avgx.cpp Normal file
View File

@ -0,0 +1,270 @@
/* Copyright (C) 2017 MariaDB Corporaton
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; version 2 of
the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
MA 02110-1301, USA. */
#include <sstream>
#include <cstring>
#include <typeinfo>
#include "regr_avgx.h"
#include "bytestream.h"
#include "objectreader.h"
using namespace mcsv1sdk;
#define DATATYPE double
// Use the simple data model
struct regr_avgx_data
{
double sum;
uint64_t cnt;
};
mcsv1_UDAF::ReturnCode regr_avgx::init(mcsv1Context* context,
ColumnDatum* colTypes)
{
if (context->getParameterCount() != 2)
{
// The error message will be prepended with
// "The storage engine for the table doesn't support "
context->setErrorMessage("regr_avgx() with other than 2 arguments");
return mcsv1_UDAF::ERROR;
}
if (!(isNumeric(colTypes[1].dataType)))
{
// The error message will be prepended with
// "The storage engine for the table doesn't support "
context->setErrorMessage("regr_avgx() with a non-numeric x argument");
return mcsv1_UDAF::ERROR;
}
context->setUserDataSize(sizeof(regr_avgx_data));
context->setResultType(CalpontSystemCatalog::DOUBLE);
context->setColWidth(8);
context->setScale(colTypes[1].scale + 4);
context->setPrecision(19);
context->setRunFlag(mcsv1sdk::UDAF_IGNORE_NULLS);
return mcsv1_UDAF::SUCCESS;
}
mcsv1_UDAF::ReturnCode regr_avgx::reset(mcsv1Context* context)
{
struct regr_avgx_data* data = (struct regr_avgx_data*)context->getUserData()->data;
data->sum = 0;
data->cnt = 0;
return mcsv1_UDAF::SUCCESS;
}
mcsv1_UDAF::ReturnCode regr_avgx::nextValue(mcsv1Context* context, ColumnDatum* valsIn)
{
static_any::any& valIn_y = valsIn[0].columnData;
static_any::any& valIn_x = valsIn[1].columnData;
struct regr_avgx_data* data = (struct regr_avgx_data*)context->getUserData()->data;
DATATYPE val = 0.0;
if (context->isParamNull(0) || context->isParamNull(1))
{
return mcsv1_UDAF::SUCCESS; // Ought not happen when UDAF_IGNORE_NULLS is on.
}
if (valIn_x.empty() || valIn_y.empty())
{
return mcsv1_UDAF::SUCCESS; // Ought not happen when UDAF_IGNORE_NULLS is on.
}
if (valIn_x.compatible(longTypeId))
{
val = valIn_x.cast<long>();
}
else if (valIn_x.compatible(charTypeId))
{
val = valIn_x.cast<char>();
}
else if (valIn_x.compatible(scharTypeId))
{
val = valIn_x.cast<signed char>();
}
else if (valIn_x.compatible(shortTypeId))
{
val = valIn_x.cast<short>();
}
else if (valIn_x.compatible(intTypeId))
{
val = valIn_x.cast<int>();
}
else if (valIn_x.compatible(longTypeId))
{
val = valIn_x.cast<long>();
}
else if (valIn_x.compatible(llTypeId))
{
val = valIn_x.cast<long long>();
}
else if (valIn_x.compatible(ucharTypeId))
{
val = valIn_x.cast<unsigned char>();
}
else if (valIn_x.compatible(ushortTypeId))
{
val = valIn_x.cast<unsigned short>();
}
else if (valIn_x.compatible(uintTypeId))
{
val = valIn_x.cast<unsigned int>();
}
else if (valIn_x.compatible(ulongTypeId))
{
val = valIn_x.cast<unsigned long>();
}
else if (valIn_x.compatible(ullTypeId))
{
val = valIn_x.cast<unsigned long long>();
}
else if (valIn_x.compatible(floatTypeId))
{
val = valIn_x.cast<float>();
}
else if (valIn_x.compatible(doubleTypeId))
{
val = valIn_x.cast<double>();
}
// For decimal types, we need to move the decimal point.
uint32_t scale = valsIn[1].scale;
if (val != 0 && scale > 0)
{
val /= pow(10.0, (double)scale);
}
data->sum += val;
++data->cnt;
return mcsv1_UDAF::SUCCESS;
}
mcsv1_UDAF::ReturnCode regr_avgx::subEvaluate(mcsv1Context* context, const UserData* userDataIn)
{
if (!userDataIn)
{
return mcsv1_UDAF::SUCCESS;
}
struct regr_avgx_data* outData = (struct regr_avgx_data*)context->getUserData()->data;
struct regr_avgx_data* inData = (struct regr_avgx_data*)userDataIn->data;
outData->sum += inData->sum;
outData->cnt += inData->cnt;
return mcsv1_UDAF::SUCCESS;
}
mcsv1_UDAF::ReturnCode regr_avgx::evaluate(mcsv1Context* context, static_any::any& valOut)
{
struct regr_avgx_data* data = (struct regr_avgx_data*)context->getUserData()->data;
if (data->cnt == 0)
{
valOut = 0;
}
else
{
valOut = data->sum / (double)data->cnt;
}
return mcsv1_UDAF::SUCCESS;
}
mcsv1_UDAF::ReturnCode regr_avgx::dropValue(mcsv1Context* context, ColumnDatum* valsDropped)
{
static_any::any& valIn_y = valsDropped[0].columnData;
static_any::any& valIn_x = valsDropped[1].columnData;
struct regr_avgx_data* data = (struct regr_avgx_data*)context->getUserData()->data;
DATATYPE val = 0.0;
if (valIn_x.empty() || valIn_y.empty())
{
return mcsv1_UDAF::SUCCESS; // Ought not happen when UDAF_IGNORE_NULLS is on.
}
if (valIn_x.compatible(charTypeId))
{
val = valIn_x.cast<char>();
}
else if (valIn_x.compatible(scharTypeId))
{
val = valIn_x.cast<signed char>();
}
else if (valIn_x.compatible(shortTypeId))
{
val = valIn_x.cast<short>();
}
else if (valIn_x.compatible(intTypeId))
{
val = valIn_x.cast<int>();
}
else if (valIn_x.compatible(longTypeId))
{
val = valIn_x.cast<long>();
}
else if (valIn_x.compatible(llTypeId))
{
val = valIn_x.cast<long long>();
}
else if (valIn_x.compatible(ucharTypeId))
{
val = valIn_x.cast<unsigned char>();
}
else if (valIn_x.compatible(ushortTypeId))
{
val = valIn_x.cast<unsigned short>();
}
else if (valIn_x.compatible(uintTypeId))
{
val = valIn_x.cast<unsigned int>();
}
else if (valIn_x.compatible(ulongTypeId))
{
val = valIn_x.cast<unsigned long>();
}
else if (valIn_x.compatible(ullTypeId))
{
val = valIn_x.cast<unsigned long long>();
}
else if (valIn_x.compatible(floatTypeId))
{
val = valIn_x.cast<float>();
}
else if (valIn_x.compatible(doubleTypeId))
{
val = valIn_x.cast<double>();
}
// For decimal types, we need to move the decimal point.
uint32_t scale = valsDropped[1].scale;
if (val != 0 && scale > 0)
{
val /= pow(10.0, (double)scale);
}
data->sum -= val;
--data->cnt;
return mcsv1_UDAF::SUCCESS;
}

99
utils/udfsdk/regr_avgx.h Normal file
View File

@ -0,0 +1,99 @@
/* Copyright (C) 2017 MariaDB Corporaton
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; version 2 of
the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
MA 02110-1301, USA. */
/***********************************************************************
* $Id$
*
* regr_avgx.h
***********************************************************************/
/**
* Columnstore interface for for the regr_avgx function
*
*
* CREATE AGGREGATE FUNCTION regr_avgx returns REAL soname
* 'libudf_mysql.so';
*
*/
#ifndef HEADER_regr_avgx
#define HEADER_regr_avgx
#include <cstdlib>
#include <string>
#include <vector>
#include <boost/any.hpp>
#ifdef _MSC_VER
#include <unordered_map>
#else
#include <tr1/unordered_map>
#endif
#include "mcsv1_udaf.h"
#include "calpontsystemcatalog.h"
#include "windowfunctioncolumn.h"
using namespace execplan;
#if defined(_MSC_VER) && defined(xxxRGNODE_DLLEXPORT)
#define EXPORT __declspec(dllexport)
#else
#define EXPORT
#endif
namespace mcsv1sdk
{
// Override mcsv1_UDAF to build your User Defined Aggregate (UDAF) and/or
// User Defined Analytic Function (UDAnF).
// These will be singleton classes, so don't put any instance
// specific data in here. All instance data is stored in mcsv1Context
// passed to each user function and retrieved by the getUserData() method.
//
// Each API function returns a ReturnCode. If ERROR is returned at any time,
// the query is aborted, getInterrupted() will begin to return true and the
// message set in config->setErrorMessage() is returned to MariaDB.
// Return the regr_avgx value of the dataset
class regr_avgx : public mcsv1_UDAF
{
public:
// Defaults OK
regr_avgx() : mcsv1_UDAF() {};
virtual ~regr_avgx() {};
virtual ReturnCode init(mcsv1Context* context,
ColumnDatum* colTypes);
virtual ReturnCode reset(mcsv1Context* context);
virtual ReturnCode nextValue(mcsv1Context* context, ColumnDatum* valsIn);
virtual ReturnCode subEvaluate(mcsv1Context* context, const UserData* valIn);
virtual ReturnCode evaluate(mcsv1Context* context, static_any::any& valOut);
virtual ReturnCode dropValue(mcsv1Context* context, ColumnDatum* valsDropped);
protected:
};
}; // namespace
#undef EXPORT
#endif // HEADER_regr_avgx.h