Home/ Questions/Q 8294527
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-06-08T14:14:53+00:00

I’m looking for a way to create vertical tables in SAS where the variables

  • 0

I’m looking for a way to create vertical tables in SAS where the variables are each treated as rows (as opposed to each row being an observation).

For example lets say I have some data for a bunch of companies, some of which is more important than others. It is easy to make proc report spit out a summary table with a few variables like this:

Name Price Shares MarketCap
co1    $5    100    $500
co2    $1    100    $100
co3    $2    200    $400

What I want to do after this is print a page of detailed information for each company which is essentially a table with a column for the description and a column for the value (and maybe a third column for the calculation).

Company 1

   Location:   CA
        CEO:   Bob Johnson
   Industry:   Semiconductors

     Shares:   100
Share Price:   $5
 Market Cap:   $500

The only way I can think of to do this in SAS is to basically transpose everything, create a new character variable that has the label (Location, Stock Price, Etc) and a second character variable that has the value and then make a two column report BY company to get a page for each. This is messy since some of the values are numeric and others are character so to get them to display on one column requires creating a new character variable and filling it with text versions of the numeric variables.

I figure there has got to be an easier way to create a vertical table since there are so many easy ways to create the horizontal tables.

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    There is also this solution which is probably better for your needs.

    First create a HTML file that will be used as a template. Wherever you want to put a value, use a macro variable as a placeholder like so:

    <html>
<h1> My title is &title </h1><br>
Name: &name <br>
Value of Blah: &blah
</html>

    Make it as attractive looking as you like.

    Next create a macro that will import the HTML template, replace the placeholders with actual values and save the result to a new file:

    /*****************************************************************************
**  PROGRAM: MACRO.RESOLVE_FILE.SAS
**
**  READS IN A FILE AND REPLACES ANY MACRO REFERENCES IN THE FILE WITH THE 
**  ACTUAL MACRO VALUES.  EG.  IF THE FILE WAS AN HTML FILE AND IT CONTAINED 
**  THE FOLLOWING HTML:
**
**    <TITLE>&HTML_TITLE</TITLE>
**
**  THEN THE PROGRAM WOULD READ THE FILE IN AND RESOLVE IT SO THAT THE OUTPUT
**  LOOKED LIKE THIS:
**
**    <TITLE>ROB</TITLE>
**  
**  ... WHEN THE MACRO VARIABLE "HTML_TITLE" EXISTED AND CONTAINED A VALUE OF 
**  "ROB".  THIS IS USEFUL WHEN YOU NEED TO CREATE "DYNAMIC" HTML FILES FROM 
**  SAS BUT DONT WANT TO DO IT FROM A DATASTEP USING PUT STATEMENTS.  DOING
**  IT THIS WAY IS MUCH CLEANER.
**
**  PARAMETERS: NONE
**
******************************************************************************
**  HISTORY:
**  1.0 MODIFIED: 22-JUL-2010  BY:RP
**  - CREATED. 
**  1.1 MODIFIED: 18-FEB-2011  BY:RP
**  - ADDED LRECL OF 32K TO STOP TRUNCATION
*****************************************************************************/
%macro resolve_file(iFileIn=, iFileOut=);
  data _null_;
    length line $32767;
    infile "&iFileIn" truncover lrecl=32767;
    file   "&iFileOut" lrecl=32767;
    input; 
    line = resolve(_infile_);
    len = length(line);
    put line $varying. len;
  run;
%mend;

    Create some test data. Also create some commands to call the above macro and pass in the values from the dataset:

    data mydata;
  attrib name length=$10 format=$10.    label='FirstName'
         blah length=6   format=comma6. label='SomeValue'
         cmd1  length=$1000
         cmd2  length=$1000
         ;

  title = 1; 
  name = "Rob" ; 
  blah = 1000; 
  cmd1 = cats('%let title=',title,';',
              '%let name=',name,';',
              '%let blah=',blah,';');
  cmd2 = cats('%resolve_file(iFileIn=c:\template.html, iFileOut=c:\result',title,'.html);');
  output;

  title = 2; 
  name = "Pete"; 
  blah = 100 ; 
  cmd1 = cats('%let title=',title,';',
              '%let name=',name,';',
              '%let blah=',blah,';');
  cmd2 = cats('%resolve_file(iFileIn=c:\template.html, iFileOut=c:\result',title,'.html);');
  output;
run;

    Use call execute to run the cmd1 and cmd2 that we created in the prior dataset. We have to only execute call execute on 1 row at a time so that the correct macro variables are used so do it using a loop. First calculate the number of rows in your dataset using your preferred technique:

    proc sql noprint;      
  select count(*) into :nobs from mydata;
quit;

    Then iterate through the dataset executing the commands one at a time and building each row to a new file:

    %macro publish;
  %local tmp;
  %do tmp = 1 %to &nobs;
    data _null_;
      set mydata(firstobs=&tmp obs=&tmp);
      call execute (cmd1);
      call execute (cmd2);
    run;
  %end;
%mend;
%publish;

    That should do the trick.

    • 0