SAS关于宏、宏函数、宏变量、data步、proc步和call execute的理解

SAS宏和宏函数的问题困扰了我三年之久，终于在昨日想通了，而想通的原因也很搞笑，仅仅是当时意识到了sas宏和宏函数是两个东西，自定的宏并不是宏函数（在其他编程语言中自定义函数和语言本身函数是一样的，受此影响！）

转载请注明出处：http://www.cnblogs.com/SSSR/p/6380957.html

结论：sas宏并不具备sas宏函数的功能，它仅仅只是一段文本，这段文本中如果有参数和宏函数，我们只是把参数替换掉和宏函数执行了，然后生成一个正常的文本（包含data步和proc步）提交给sas运行。

遇到宏函数时会直接执行，遇到宏时会直接进行文本替换（宏中的宏函数也会直接执行），宏函数返回的文本会和其他的data步和proc步组合然后一起提交给sas步运行。

宏函数直接执行，sas函数需要在data步中执行（%put和put的区别）

下面来看两个简单的例子：

options symbolgen mprint;*将宏编译的过程也打印出来，以便进行测试;
%macro s(i);

    %put &i.;
%mend;

data _null_;
     set sashelp.class;
     put name;
     %s('program');
run;


/*以下是结果：汉字为注释

105  %macro s(i);
106
107      %put &i.;
108  %mend;
109
110  data _null_;
111  set sashelp.class;
112  put name;
113       %s('program');                             //下面是这句宏调用的解析
SYMBOLGEN:  Macro variable I resolves to 'program'  // I是%s的宏参数，被替换成了program。下面是直接打印了program，这个是为啥呢？就是因为%put是宏函数，会立即执行，从而先打印了program，然后才会将正常的data步提交给sas执行，run;后的部分是宏编译完以后执行的结果。
'program'
114  run;

Alfred
Alice
Barbara
Carol
Henry
James
Jane
Janet
Jeffrey
John
Joyce
Judy
Louise
Mary
Philip
Robert
Ronald
Thomas
William
NOTE: There were 19 observations read from the data set SASHELP.CLASS.
NOTE: DATA statement used (Total process time):
      real time           0.01 seconds


*/

%macro t(i);

    put &i.;
%mend;

data _null_;
set sashelp.class;
     %t(name);
	 %s('programe')
run;
/*以下为结果：中文为注释

135  %macro t(i);
136
137      put &i.;
138  %mend;
139
140  data _null_;
141  set sashelp.class;
142       %t(name);
SYMBOLGEN:  Macro variable I resolves to name    //宏变量I被解析成name，注意这个name不带双引号，
MPRINT(T):   put name;                           //这里put name和data set等一起被提交给sas执行
143       %s('programe')
SYMBOLGEN:  Macro variable I resolves to 'programe'  //这个因为是%put所以被直接执行了。
'programe'
144  run;

Alfred
Alice
Barbara
Carol
Henry
James
Jane
Janet
Jeffrey
John
Joyce
Judy
Louise
Mary
Philip
Robert
Ronald
Thomas
William
NOTE: There were 19 observations read from the data set SASHELP.CLASS.
NOTE: DATA statement used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds


*/



proc print data=sashelp.class;
run;

　　

 再来看一个复杂的例子：

options symbolgen mprint;
data b;
input a1 a2 a3;
cards;
1 2 3
;
run;

%macro varx;
  %do i=0 %to 2;
    proc sql  ;
      select a%eval(&i.+1) into:a%eval(&i.+1)
      from b;
    quit;
   %let j=%eval(&i.+1);
   data _null_;
      a= &&a&j.;
      put a;
   run;
%end ;
%mend varx;
%varx;

/*以下为结果：中文为注释，整个程序循环了三次i=0 1 2,所以会生成三次proc sql和三次data步，然后提交给sas运行。从这个就可以看出sas宏的一个作用，减少代码量，将重复的代码进行封装。
145  data b;
146  input a1 a2 a3;
147  cards;

NOTE: The data set WORK.B has 1 observations and 3 variables.
NOTE: DATA statement used (Total process time):
      real time           0.01 seconds
      cpu time            0.01 seconds


149  ;
150  run;               //以上为数据集生成
151
152  %macro varx;
153  %do i=0 %to 2;
154  proc sql  ;
155  select a%eval(&i.+1) into:a%eval(&i.+1)
156  from b;
157  quit;
158  %let j=%eval(&i.+1);
159  data _null_;
160  a= &&a&j.;
161  put a;
162  run;
163  %end ;
164  %mend varx;
165  %varx;
MPRINT(VARX):   proc sql ;
SYMBOLGEN:  Macro variable I resolves to 0
SYMBOLGEN:  Macro variable I resolves to 0
MPRINT(VARX):   select a1 into:a1 from b;
MPRINT(VARX):   quit;                                      //从proc sql到这里就是宏varx编译生成的一段proc sql代码，这部分提交给sas运行。
NOTE: PROCEDURE SQL used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds


SYMBOLGEN:  Macro variable I resolves to 0
MPRINT(VARX):   data _null_;
SYMBOLGEN:  && resolves to &.
SYMBOLGEN:  Macro variable J resolves to 1
SYMBOLGEN:  Macro variable A1 resolves to        1
MPRINT(VARX):   a= 1;
MPRINT(VARX):   put a;
MPRINT(VARX):   run;            //从data步到这一行是宏varx编译生成的第二段代码data步，这段代码中宏变量I被编译成了0，j直接被计算成了1.这段代码也同样展示了select into生成的宏变量在其他程序中调用的执行顺序。

1
NOTE: DATA statement used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds


MPRINT(VARX):   proc sql ;
SYMBOLGEN:  Macro variable I resolves to 1
SYMBOLGEN:  Macro variable I resolves to 1
MPRINT(VARX):   select a2 into:a2 from b;
MPRINT(VARX):   quit;               //重复proc sql
NOTE: PROCEDURE SQL used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds


SYMBOLGEN:  Macro variable I resolves to 1
MPRINT(VARX):   data _null_;
SYMBOLGEN:  && resolves to &.
SYMBOLGEN:  Macro variable J resolves to 2
SYMBOLGEN:  Macro variable A2 resolves to        2
MPRINT(VARX):   a= 2;
MPRINT(VARX):   put a;
MPRINT(VARX):   run;                      //重复data 

2
NOTE: DATA statement used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds


MPRINT(VARX):   proc sql ;
SYMBOLGEN:  Macro variable I resolves to 2
SYMBOLGEN:  Macro variable I resolves to 2
MPRINT(VARX):   select a3 into:a3 from b;
MPRINT(VARX):   quit;                         //重复proc sql
NOTE: PROCEDURE SQL used (Total process time):
      real time           0.01 seconds
      cpu time            0.01 seconds


SYMBOLGEN:  Macro variable I resolves to 2
MPRINT(VARX):   data _null_;
SYMBOLGEN:  && resolves to &.
SYMBOLGEN:  Macro variable J resolves to 3
SYMBOLGEN:  Macro variable A3 resolves to        3
MPRINT(VARX):   a= 3;
MPRINT(VARX):   put a; 
MPRINT(VARX):   run;                     //重复data

3
NOTE: DATA statement used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds


*/

　　未完，后续预告call execute（http://bbs.pinggu.org/thread-2377205-1-1.html 此链接的讲解非常透彻，同时call excute可以调用宏，而且讲解了在使用宏和宏变量时单引号和双引号的处理）。

在Python和R语言中，有一个for循环的功能非常好用，但是在sas中没有类似的功能。鉴于sas data步中的pdv是一行一行的读取执行的，so可以结合call execute来实现for循环功能：

以下代码实现的功能是将三行数据生成三个数据集，数据集的名称是dataset列的值。

data a;

   input A B dataset $;

   datalines;

1 2 ds_1

6 7 ds_2a

2 3 ds_100c

;

run;

data _null_;

  set a;

  call execute('data ' ||dataset||';'||'set  a (firstobs=' ||_n_|| ' obs=' || _n_ || ');run;' );

run;

/* or */

data _null_;

  set a;

  call execute('data ' || dataset || ';' || 'set  a ; if  dataset= "' || dataset || '"; run ;' );

run;