topic Re: need to replace the varible names with the values of the first observation. in SAS Procedures

need to replace the varible names with the values of the first observation.

need_sas_help — Fri, 03 Oct 2014 15:43:26 GMT

I have the following Datatset with all characrter variables (SS,DD,YY,PP,NN):

SS     DD    YY      PP       NN
--    ---    --     ---       --
hy     tr    uy      io       oi
fr      1     3       5        6
er      4     5       7       10

I need to replace the varible names with the values of the first observation.
So the new variable names will be (hy, tr, uy, io, oi) can't do any hard coding.
the result dataset should be as below:

hy     tr    uy      io       oi
--    ---    --     ---       --
fr      1     3       5        6
er      4     5       7       10

Thank you so much from the bottom of my heart for helping me.

Re: need to replace the varible names with the values of the first observation.

Reeza — Fri, 03 Oct 2014 15:58:59 GMT

How did you get this data, the best place would be to change how you imported it in.

That being said, is this one time, with only 5 columns or it could be more?

Re: need to replace the varible names with the values of the first observation.

Reeza — Fri, 03 Oct 2014 16:12:37 GMT

This answers your question, but I still think you should fix your import, especially since all your variables are character when you probably want them to be numeric:

data have;

input SS $ DD $ YY $ PP $ NN $;

cards;

hy tr uy io oi

fr 1 3 5 6

er 4 5 7 10

;

run;

proc transpose data=have out=names;

var _char_;

run;

proc sql;

select catx(" = ", _name_ , col1)

into :rename_list separated by " "

from names;

quit;

%put &rename_list;

data want;

set have (firstobs=2);

rename &rename_list;

run;

Re: need to replace the varible names with the values of the first observation.

Ksharp — Sat, 04 Oct 2014 11:19:40 GMT

As Reeza pointed out .You should fix the importing way . Not consider change its name.

data have;
input SS $ DD $ YY $ PP  $ NN $;
cards;
hy     tr    uy      io       oi
fr      1     3       5        6
er      4     5       7       10
;
run;
proc sql noprint;
select name into : list separated by ' ' 
 from dictionary.columns
  where libname='WORK' and memname='HAVE' ;
quit;
data _null_;
 set have(obs=1);
 length x y $ 32767;
 x=catx(' ',of _all_);
 do i=1 to countw(x);
  y=catx(' ',y,catx('=',scan("&list",i),scan(x,i)));
 end;
 call symputx('rename',y);
run;
data want;
  set have (firstobs=2);
  rename &rename ;
run;

Xia Keshan

Re: need to replace the varible names with the values of the first observation.

Tom — Sat, 04 Oct 2014 15:55:18 GMT

Add OBS=1 dataset option to the input to the PROC TRANSPOSE to prevent it from transposing the whole dataset.

Re: need to replace the varible names with the values of the first observation.

Loko — Mon, 06 Oct 2014 17:41:40 GMT

Hello,

data have;

input SS $ DD $ YY $ PP $ NN $;

datalines;

hy tr uy io oi

fr 1 3 5 6

er 4 5 7 10

;

data _null_;

set have (obs=1);

array allvar{*} ss--nn ;

length syntax $ 50;

do i=1 to dim(allvar);

syntax=strip(syntax) || ' ' || catt(vname(allvar{i}),'=',allvar{i});

put syntax=;

call symputx ('newvars', 'rename ' || strip (syntax) );

end;

run;

data want;

set have (firstobs=2);

&newvars;

run;