transpose in one step

[econfkw]

The original data is test22.

1. I'm using 9.1, always have problem with informat and format date. By import, SAS reads it as character.

2. How can I transpose test22 to test24.

thanks

[Tom]

What that you want to "transpose"?

Do the empty rows at the end of test22.csv have any meaning?

Looks like you want to do:

data test24; set test22;

  s=1 ; return=s1; output;

  s=2 ; return=s2; output;

  drop s1 s2;

run;

test22.csv

---

Date,s1,s2,premium,smb

1952Q1,-1.171800189,-1.77499335,3.22,-2.74

1952Q2,-10.25698639,-1.812109541,1.85,-2.08

1952Q3,-3.994618175,0.55753206,-1.9,1.96

1952Q4,-1.262761712,2.255299066,8.34,-3.32

,,,,

,,,,

,,,,

,,,,

---

test24.csv

---

Date,s,return,premium,smb

1952Q1,1,-1.171800189,3.22,-2.74

1952Q1,2,-1.77499335,3.22,-2.74

1952Q2,1,-10.25698639,1.85,-2.08

1952Q2,2,-1.812109541,1.85,-2.08

1952Q3,1,-3.994618175,-1.9,1.96

1952Q3,2,0.55753206,-1.9,1.96

1952Q4,1,-1.262761712,8.34,-3.32

1952Q4,2,2.255299066,8.34,-3.32

---

[econfkw]

sorry, the dots have no meaning. Your code does work. I have s1-s25, I think a do loop may work.

Anyhow, regarding question 1, how can I informat the date variable as date instead of character?

thanks

[Tom]

You can use YYQ informat and format to read those values into actual dates and redisplay them at the same level of precision.

The dates will be stored as the first day of the quarter. So 1952Q1 will be stored as '01JAN1952'd.

data test22 ;

  infile 'test22.csv' dsd dlm=',' truncover firstobs=2 ;

  input date s1 s2 premium smb ;

  informat date yyq6.;

  format date yyq6.;

run;