[原創]擴展求方差的mysql函數例子

原創

2020-02-22 05:30

源碼

// Author: JiangMiao
// Name: 擴展求方差的mysql函數列子
// Date: 2006-10-19
// Link: http://blog.csdn.net/antter - JiangMiao的Blog
#include "winsock2.h"
#include "mysql.h"
#include <vector>
using namespace std;
#define SAFE_DELETE(p) if(p!=NULL){delete p;p=NULL;}
#define CDLLEXPORT extern "C" __declspec(dllexport)
typedef __int64 longlong;
typedef vector<double> vec_double;
typedef unsigned long ulong;
class VAR
{
private:
  vec_double datas;
  double total;
public:
  VAR():total(0.0) {}
  //加入num
  void push_back(double num)
   {
   datas.push_back(num);
   total+=num;
   }
  void clear()
   {
   datas.clear();
   total=0.0;
   }
  //取方差
  double getVariance()
   {
   size_t count=datas.size();
   double avr=0.0;
   avr=(total/count); //平均數
   double rt=0.0;
   for(size_t i=0;i<count;i++)
    {
    double k=(datas[i]-avr);
    rt+=k*k;
    }
   return rt/count;
   }

};
CDLLEXPORT my_bool variance_init(UDF_INIT *initid, UDF_ARGS *args, char *message)
{
initid->ptr = NULL;
if(args->arg_count!=1) //參數個數爲1
  {
  return 1;
  }
if(args->arg_type[0]!=REAL_RESULT||args->arg_type[0]!=INT_RESULT) //參數類別爲整型或double
  {
  return 1;
  }
initid->ptr = (char*)new VAR();
return 0;
}
CDLLEXPORT void variance_deinit(UDF_INIT *initid)
{
VAR* ptr=(VAR*)initid->ptr;
delete ptr;
}
CDLLEXPORT double variance(UDF_INIT *initid, UDF_ARGS *args,char *is_null, char *error)
{
VAR* ptr=(VAR*)initid->ptr;
return ptr->getVariance();
}
CDLLEXPORT void variance_clear(UDF_INIT *initid, char *is_null, char *error)
{
VAR* ptr=(VAR*)initid->ptr;
ptr->clear();
}
CDLLEXPORT void variance_add(UDF_INIT *initid, UDF_ARGS *args, char *is_null, char *error)
{
VAR* ptr=(VAR*)initid->ptr;
char* argo=args->args[0];
double arg;
if(args->arg_type[0]==REAL_RESULT)
  {
  arg=*(double*)argo;
  }
if(args->arg_type[0]==INT_RESULT)
  {
  arg=(double)*(__int64*)argo;
  }
ptr->push_back(arg);
}

編譯後得到variance.dll

複製到bin目錄下

測試

mysql> use test;
Database changed
mysql> create table vartest (realtest real,inttest int);
Query OK, 0 rows affected (0.11 sec)

mysql> insert into vartest values(5,5),(6,6),(9,9),(10,10),(5,5);
Query OK, 5 rows affected (0.03 sec)
Records: 5 Duplicates: 0 Warnings: 0

mysql> create aggregate function variance returns real soname 'variance.dll';
Query OK, 0 rows affected (0.00 sec)

mysql> select variance(realtest),variance(inttest) from vartest;
+--------------------+-------------------+
| variance(realtest) | variance(inttest) |
+--------------------+-------------------+
| 4.4 | 4.4000 |
+--------------------+-------------------+
1 row in set (0.00 sec)

mysql>

如果有建議或疑問歡迎留言

我的blog是http://blog.csdn.net/antter