How can I make a global change to all the blank values to the actual correct value within a set of variables?

VCATALG	Library name, Member name, Member type, Object name, Object type, Object description, Date modified, Object alias.
VCOLUMN	Library name, Member name, Member type, Column name, Column type, Column length, Column position, Column number, Column label, Column format, Column informat, Column index type
VEXTFL	Fileref, Path name (fully qualified file name and path), Engine name.
VMEMBER	Library name, Member name, Member type, Engine name, Indexes, Path name.
VOPTION	Session option name, Session option setting, Option description.
VSCATLG	Library name, Member name.
VSLIB	Library name, Path name
VSTABLE	Library name, Member name
VTABLE	Library name, Member name, Member type, Dataset label, Date created, Date modified, Number of observations, Observation length, Number of variables, Type of password protection, Compression routine, Reuse space, Bufsize, Number of deleted observations, Type of indexes.
VTITLE	Title number, Title text.

This example prints a list of the libraries currently assigned on your system.

proc print data = sashelp.vslib;
run;

Step 1
Sometimes you are presented with data that contains blank values for character fields. However in reality, these blank values may have a value. In the example below we create a dataset containing details of current account balances. Zero balances are currently represented with blanks, but SAS will treat these as missing values. We will use SAS views to build a macro that will replace all instances of these missing values with their 'true' values. If you have many variables that need values replacing (not just blank or character) a macro can prove to be more maintainable than tailor written SAS code.

data balances;
input jan03 $9. feb03 $9. mar03 $9. apr03 $9. may03 $9. jun03 $9.;
cards;
OverdrawnOverdrawnOverdrawnOverdrawnOverdrawnOverdrawn
Credit            Credit   Credit   Overdrawn
OverdrawnCredit   Overdrawn         OverdrawnCredit
         OverdrawnOverdrawn         Credit   Credit
;
run;

proc print data=balances;
run;

Step 2
This is where we use a SAS view, in this instance sashelp.volumn to populate the view tmpz1_char1 (remember that views take up less space than datasets) with details of the char variable names held within the WORK.BALANCES data set. Notice in the SQL, we use uppercase letters for the library and dataset name. This is because the view creates uppercase values for the libname and memname fields.

proc sql;
create view tmpz1_char1 as select * from sashelp.vcolumn
where libname='WORK' and memname='BALANCES' and type='char';
quit;

data _null_;
set tmpz1_char1 end=eof;
call symput('charz'||left(put(_n_,8.)),(trim(name)));
if eof then call symput('count',put(_n_,8.));
run;

%macro testdatz (indata,outdata,replacestr,withstr);

We use 4 positional parameters for this macro. The first 'indata', will be used to tell us which dataset we wish to replace the values in. The second is 'outdata' which is the name given to the resulting dataset created after the replacements have been made. The 'replacestr' parameter is the string that you would like to replace and finally 'withstr' is the character value that replaces existing values.

%local i; - will be used to generate a loop
data &outdata; - create resulting dataset
set &indata; - access the input dataset
%do i=1 %to &count; - use the count macro variable to define the number of loops
if &&charz&i="&replacestr" then &&charz&i="&withstr"; - here's where we replace the values within the dataset
%end;
run;

This is the end of the macro.

%mend;

Step 5
Here's where we run the macro. Notice the values of our conditional parameters

%testdatz (work.balances,work.balances1, ,Zero);

Step 6
Lets print the resulting dataset work.balances1 to see the replaced values.

proc print data=balances1;
run;

Note: The above code has been tested using SAS 8.2 on the Windows operating system.