ok，自己弄明白了
array _have (3) xys xjz xms; 应该不能用array _have (3)  xms-xys这个句子
call sortn(value3,value2,value1);也不能写为call sortn(value3-value1);
原因不懂。。。。

array _have (3) xys xjz xms;  可以用array _have (3) xys--xms;
中间的两个‘-’表示按照变量在数据集中的位置从xys到xms的所有变量。

真的是学习了 非常感谢

Unless the performance is an issue, saving a couple of seconds is not much. Simple to understanding and maintenance free are much more important in programming.

Still in this case, the transpose beats the other one by about factor 3. See the log attached below.

157  data test;
158    array x(*) x1-x100;
159    do date=1 to 10000;
160       do j=1 to dim(x);
161         x[j]=ranuni(123);
162       end;
163       output;
164    end;
165    keep date x:;
166  run;

NOTE: The data set WORK.TEST has 10000 observations and 101 variables.
NOTE: DATA statement used (Total process time):
      real time           0.05 seconds
      cpu time            0.06 seconds


167
168  %let time=%sysfunc(time( ));
169  data a(keep= date value:) b(keep=date var:) ;
170     set test;
171     array _have (*) x1-x100;
172     array _sorted (*) value1-value100;
173     array _name (*) $ var1-var100;
174
175     *** Sort values;
176     do i=1 to dim(_have);
177        _sorted(i)=_have(i);
178     end;
179     call sortn(value3,value2,value1);
180
181     *** Get variable names;
182     do i=1 to dim(_sorted);
183        do j=1 to dim(_Have);
184                if _have(j)=_sorted(i) then _name(i)=vname(_have(j));
185            end;
186     end;
187  run;

NOTE: There were 10000 observations read from the data set WORK.TEST.
NOTE: The data set WORK.A has 10000 observations and 101 variables.
NOTE: The data set WORK.B has 10000 observations and 101 variables.
NOTE: DATA statement used (Total process time):
      real time           3.07 seconds
      cpu time            3.07 seconds


188
189  %put >>>>>time of array= %sysevalf(%sysfunc(time( ))-&time)<<<<<;
>>>>>time of array= 3.0779999998922<<<<<
190
191  %let time=%sysfunc(time( ));
192
193  proc transpose data=test out=test2;
194  by date;
195  var x1-x100;
196  run;

NOTE: There were 10000 observations read from the data set WORK.TEST.
NOTE: The data set WORK.TEST2 has 1000000 observations and 3 variables.
NOTE: PROCEDURE TRANSPOSE used (Total process time):
      real time           0.20 seconds
      cpu time            0.18 seconds


197
198  proc sort data=test2;
199  by date decending col1 ;
200  run;

NOTE: There were 1000000 observations read from the data set WORK.TEST2.
NOTE: The data set WORK.TEST2 has 1000000 observations and 3 variables.
NOTE: PROCEDURE SORT used (Total process time):
      real time           0.31 seconds
      cpu time            0.73 seconds


201
202  proc transpose data=test2(drop=_name_) out=a(drop=_name_) prefix=value;
203  by date;
204  var  col1;
205  run;

NOTE: There were 1000000 observations read from the data set WORK.TEST2.
NOTE: The data set WORK.A has 10000 observations and 101 variables.
NOTE: PROCEDURE TRANSPOSE used (Total process time):
      real time           0.22 seconds
      cpu time            0.20 seconds


206
207
208  proc transpose data=test2 out=b(drop=_:) prefix=namevar;
209  by date;
210  var _name_ ;
211  run;

NOTE: There were 1000000 observations read from the data set WORK.TEST2.
NOTE: The data set WORK.B has 10000 observations and 101 variables.
NOTE: PROCEDURE TRANSPOSE used (Total process time):
      real time           0.36 seconds
      cpu time            0.23 seconds


212
213  %put >>>>>time of transpose= %sysevalf(%sysfunc(time( ))-&time)<<<<<;
>>>>>time of transpose= 1.13199999999778<<<<<