SAS中观测内变量通用拼接：保留数值/货币显示格式的方案问询

Great question! The problem with using CATX() directly is that it converts numeric variables to their raw underlying values, stripping away any formatting like currency symbols, decimal places, or date masks. Here's a fully generalizable solution that retains the exact displayed format of every variable (whether character, numeric, currency, date, etc.) when concatenating them into a single field:

Step 1: Generate Formatted Conversion Expressions #

First, we'll query SAS's metadata to create a list of expressions that convert each variable to its formatted character representation. This handles both character and numeric variables appropriately:

proc sql noprint;
  select 
    case 
      /* For numeric variables: use PUT() with their assigned format (or BEST12. as fallback) */
      when type = 'num' then catx(' ', 'put(', name, ',', cats(coalesce(format, 'best12.'), '))')
      /* For character variables: just use the variable name directly */
      else name
    end as put_expr
  into :put_list separated by ','
  from dictionary.columns
  where libname = 'SASHELP' and memname = 'SHOES' /* Replace with your library/dataset */
  order by varnum; /* Keep variables in their original dataset order */
quit;

Step 2: Concatenate Formatted Values #

Next, use the generated &put_list. macro variable to concatenate all formatted values into a single variable:

data stuff;
  set sashelp.shoes (obs=10); /* Replace with your target dataset */
  length all $5000; /* Adjust this length based on your dataset's total character needs */
  all = catx(',', &put_list.); /* CATX adds commas and automatically skips missing values */
run;

/* Verify the formatted output */
proc print data=stuff;
  var all;
run;

How This Works #

• Metadata-Driven Logic: The dictionary.columns table gives us critical details about every variable in the dataset—its type (character/numeric) and assigned format. We build custom PUT() expressions for numeric variables to preserve their formatting, while character variables are used directly since they already store their displayed value.
• Full Generalizability: Swap out SASHELP.SHOES with any dataset, and this code adapts automatically—no manual tweaks needed for different variable types or formats.
• Clean Missing Value Handling: CATX() ensures missing values don't add extra commas to the final string, keeping the output neat and readable.

Example Output Snippet #

For the first observation in sashelp.shoes, the all variable will retain currency formatting for numeric fields, looking something like this:

Asia,Tokyo,Men's Casual,$6,682.00,$12,204.00,12,41

内容的提问来源于stack exchange，提问作者user2699504