Question

我正在尝试显示创建数据集，以显示消费者单位如何参与四个季度访谈中的至少三个。

我不确定通过查看我看到CU_ID出现多少次来使用proc freq是否有效。但到目前为止这是我的代码：

data fmli071Q1;
set hw06.fmli071;
QTR = 1;
run;

data fmli072Q2;
set hw06.fmli072;
QTR = 2;
run;

data fmli073Q3;
set hw06.fmli073;
QTR = 3;
run;

data fmli074Q4;
set hw06.fmli074;
QTR = 4;
run;

data fmli2007_bob2;
set fmli071Q1 fmli072Q2 fmli073Q3 fmli074Q4;
run;

proc contents data = fmli2007_bob2;
run;

data memi071Q1;
set hw06.memi071;
QTR = 1;
run;

data memi072Q2;
set hw06.memi072;
QTR = 2;
run;

data memi073Q3;
set hw06.memi073;
QTR = 3;
run;

data memi074Q4;
set hw06.memi074;
QTR = 4;
run;

data memi2007_bob2;
set memi071Q1 memi072Q2 memi073Q3 memi074Q4;
run;

proc sort data=fmli2007_bob2;
by CU_ID;
run;

proc sort data=memi2007_bob2;
by CU_ID;
run;

data ce2007_bob2;
merge fmli2007_bob2 memi2007_bob2
by CU_ID;
run;

现在，我需要展示在2007年的四个季度访谈中至少有三个参与了多少个消费者单位？我应该如何通过proc freq或其他方法执行此操作我想创建一个新的数据步骤调用至少_three_bob2，该调用仅显示参与2007年四个季度访谈中至少三个中的消费者单位的IDS。然后最终将其打印出来

data atleast_three_bob2;
set ce2007_bob2;
run;

Answer 1

Proc SQL可以根据复杂的标准执行数据聚合和频率计数。

data have;
  do quarter = 1 to 4;
    do id = 1 to 100;
      if quarter > 0 and mod(id,7) = 0 then 
        output;
      else
      if quarter > 1 and mod(id,13) = 0 then 
        output;
      else if quarter > 2 then 
        output;    
    end;
  end;
run;

proc sql;
  create table want(label="ids and quarters of ids appearing in >= 3 quarters") as 
  select quarter, id
  from have
  group by id
  having count(distinct quarter) >= 3
  order by id, quarter
  ;

  create table want2(label="frequency counts for number of ids having >= 3 quarters") as
  select quarter_count, count(*) as id_count 
  from
  (
    select count(*) as quarter_count
    from want
    group by id
  )
  group by quarter_count
  ;

使用Proc Freq或不同的方法

1 个答案: