Question

我有一个数据集，在不同的时间点完成了大量的实验室测试。我试图将数据集从长到大转换，但问题是实验室测试发生在不同的时间点，具体取决于测试类型。当我转置它时，我失去了分辨结果来自哪个时间点的能力。请参阅下面的示例代码：

*Create test data;
data long;
do subject=1 to 10; 
    do test=1 to 3; 
        do visit=1 to 3; 
            result=rand("Uniform");
        output; 
end; end; end;
run;

*Now remove records at certain visits depending upon the test type;
data long; set long;
    if test=2 and visit=2 then delete;
    if test=3 and visit=1 then delete;
run;

*Sort and transpose;
*Test 2 should only be at visit 1 and 3, and test 3 at visits 2 and 3;
*This transpose does not accomplish that goal;
proc sort data=long; by subject test visit;run;
proc transpose data=long out=wide;
    by subject test ;
    var result;
run;

Answer 1

对于您的示例数据，您只需要在PROC TRANSPOSE代码中添加ID语句。这样它将使用VISIT的值来命名结果列。您可能还想将PREFIX =选项添加到PROC TRANSPOSE语句中。

proc transpose data=long out=wide prefix=visit;
  by subject test ;
  id visit ;
  var result;
run;

数据丢失时SAS转置

1 个答案: