I am not entirely sure what you are after, given that you have not included an example of your desired outcome, or initial data.  I created the following some time ago, which you might find useful.  By using the ODS output you are able to concatenate the results of PROC FREQ in one simple step.  I hope it is useful in achieving your goal. DATA HAVE; INFILE DATALINES DELIMITER="," TRUNCOVER DSD; LENGTH DOCUMENTATION_VV $32. OCCUPANCY_VV $32. PROPERTY_VV $32. PURPOSE_VV $32.; INPUT DOCUMENTATION_INVALID DOCUMENTATION_VV $ OCCUPANCY_INVALID OCCUPANCY_VV $ PROPERTY_INVALID PROPERTY_VV $ PURPOSE_INVALID PURPOSE_VV $; DATALINES; 0,"FULL, SUB, VOI/VOA, NODOC, QUICK",0,"OWNER, 2NDHOME, INVESTOR",1,"SF, MF, TOWNHOUSE, CONDO",0,"PURCHASE, CASHOUT, RATE/TERM" 0,"FULL, SUB, VOI/VOA, NODOC, QUICK",1,"OWNER, 2NDHOME, INVESTOR",0,"SF, MF, TOWNHOUSE, CONDO",1,"PURCHASE, CASHOUT, RATE/TERM" 0,"FULL, SUB, VOI/VOA, NODOC, QUICK",0,"OWNER, 2NDHOME, INVESTOR",0,"SF, MF, TOWNHOUSE, CONDO",0,"PURCHASE, CASHOUT, RATE/TERM" 0,"FULL, SUB, VOI/VOA, NODOC, QUICK",0,"OWNER, 2NDHOME, INVESTOR",0,"SF, MF, TOWNHOUSE, CONDO",0,"PURCHASE, CASHOUT, RATE/TERM" 1,"FULL, SUB, VOI/VOA, NODOC, QUICK",0,"OWNER, 2NDHOME, INVESTOR",0,"SF, MF, TOWNHOUSE, CONDO",1,"PURCHASE, CASHOUT, RATE/TERM" 0,"FULL, SUB, VOI/VOA, NODOC, QUICK",1,"OWNER, 2NDHOME, INVESTOR",0,"SF, MF, TOWNHOUSE, CONDO",0,"PURCHASE, CASHOUT, RATE/TERM" ; RUN; ODS OUTPUT ONEWAYFREQS=WORK.OWF (DROP = F_: RENAME = (FREQUENCY = COUNT)); PROC FREQ DATA=HAVE;   BY DOCUMENTATION_VV OCCUPANCY_VV PROPERTY_VV PURPOSE_VV;   TABLES DOCUMENTATION_INVALID OCCUPANCY_INVALID PROPERTY_INVALID PURPOSE_INVALID /NOCUM NOPERCENT ; RUN; ODS OUTPUT CLOSE; DATA TRANSFORM (KEEP = VAL_INVAL VARIABLE VALID_VALUES DOCUMENTATION_VV COUNT);   LENGTH  VAL_INVAL $7. VARIABLE $21.;   VAL_INVAL = "VALID";   SET OWF;   BY TABLE;   IF DOCUMENTATION_INVALID = 1 THEN VAL_INVAL = "INVALID";   ELSE IF OCCUPANCY_INVALID = 1 THEN VAL_INVAL = "INVALID";   ELSE IF PROPERTY_INVALID = 1 THEN VAL_INVAL = "INVALID";   ELSE IF PURPOSE_INVALID = 1 THEN VAL_INVAL = "INVALID";   VARIABLE = SCAN(TABLE,-1);   IF VARIABLE = "DOCUMENTATION_INVALID" THEN VALID_VALUES = DOCUMENTATION_VV;   ELSE IF VARIABLE = "OCCUPANCY_INVALID" THEN VALID_VALUES = OCCUPANCY_VV;   ELSE IF VARIABLE = "PROPERTY_INVALID" THEN VALID_VALUES = PROPERTY_VV;   ELSE IF VARIABLE = "PURPOSE_INVALID" THEN VALID_VALUES = PURPOSE_VV;   OUTPUT;   IF FIRST.TABLE THEN TOTAL = COUNT;     ELSE TOTAL+COUNT;   IF LAST.TABLE  THEN DO;     COUNT = TOTAL;     VAL_INVAL = "TOTAL";     OUTPUT;   END; RUN; PROC TRANSPOSE DATA = TRANSFORM OUT=TRANS (DROP = DOCUMENTATION_VV _:) SUFFIX=_COUNT;   ID VAL_INVAL ;   BY VARIABLE VALID_VALUES DOCUMENTATION_VV NOTSORTED;     VAR COUNT; RUN; ... View more

Unfortunately there is no way to modify the length of a variable using the proc datasets modify function, however on the bright side you can use it to alter the format. data have; attrib var1 format = $15.; var1 = '1234567890'; run; proc contents data=have; run; proc datasets library=work nolist; modify have; format var1 $4.; run; proc contents data=have; run; ... View more

The following article explains the issue quite nicely. SAS(R) 9.2 Language Reference: Dictionary, Fourth Edition Let me know if you have any further questions. ... View more