data manipulation!

[River Huang]

Dear All, I was asked this question (https://bbs.pinggu.org/forum.php?mod...=1#pid57861683). Suppose that we have a dataset like

clear
input float(id indu)
1 .9375
2  .125
3     1
4  .375
5  .625
end

The purpose is to construct a new dataset like

clear
input str4 id float(id1 id2 id3 id4 id5)
"id1" .9375 .125 .9375 .375 .625
"id2"  .125 .125  .125 .125 .125
"id3" .9375 .125     1 .375 .625
"id4"  .375 .125  .375 .375 .375
"id5"  .625 .125  .625 .375 .625
end

The value for each pair, say, id1 and id2, is the minimum of id=1 (.9375 and id=2 (.125). Any suggestions to go from the data above to the data below? Thanks.

[Andrew Musau]

 forval i=1/5{
 gen id`i'= cond(indu[`i']<=indu[_n], indu[`i'], indu[_n])
 }

Result:

. l

     +-------------------------------------------------+
     | id    indu     id1    id2     id3    id4    id5 |
     |-------------------------------------------------|
  1. |  1   .9375   .9375   .125   .9375   .375   .625 |
  2. |  2    .125    .125   .125    .125   .125   .125 |
  3. |  3       1   .9375   .125       1   .375   .625 |
  4. |  4    .375    .375   .125    .375   .375   .375 |
  5. |  5    .625    .625   .125    .625   .375   .625 |
     +-------------------------------------------------+

[River Huang]

Dear Andrew, Thanks a lot for this helpful suggestion.

[Nick Cox]

Consider also as inside the loop in Andrew Musau's helpful post.

gen id`i' = min(indu[`i'], indu)