/* Windows MD5 sample code
 * Richard A. DeVenezia
 * http://www.devenezia.com
 * May 19, 2004
 */

* Sometimes you need to mask data,  MD5 is sufficient in most cases;

* One common use is storing passwords (hashed) where anyone can read them,
* access is granted only when an input password is hashed and is found to
* match an entry in the password table.;
*

* Another common use is too blind parts of data distributed to analysts.
* Typically, name, social security numbers and other identifiers that would
* de-anonymize an individual that permitted their information to be studied
* in aggregates;

filename sascbtbl catalog 'WORK.API.MD5.SOURCE';

data _null_;
  file sascbtbl;
  input;
  put _infile_;
  cards4;
routine MD5Init
   module=cryptdll
   minarg=1
   maxarg=1
   stackpop=called
;

arg 1 input num format=pib4. byvalue ; * MD5_CTX * (use addr());


routine MD5Update
   module=cryptdll
   minarg=3
   maxarg=3
   stackpop=called
;

arg 1 input num  format=pib4. byvalue ; * MD5_CTX * context (use addr());
arg 2 input char format=$cstr200. ;     * const unsigned char * bufferBeingHashed;
arg 3 input num  format=pib4. byvalue ; * unsigned int numberOfCharsInBuffer;


routine MD5Final
   module=cryptdll
   minarg=1
   maxarg=1
   stackpop=called
;

arg 1 input num format=pib4. byvalue ; * MD5_CTX * (use addr());
;;;;

run;

/* context must be large enough to contain this structure

    typedef struct {
      ULONG i[2];
      ULONG buf[4];
      unsigned char in[64];
      unsigned char digest[16];
    } MD5_CTX;

 */
data work.class;
  length name_hashed $32;
  length md5_context $110;

  set sashelp.class;

  md5_context = ' ';
  call module ('MD5Init',   addr(md5_context));
  call module ('MD5Update', addr(md5_context), name, length(name));
  call module ('MD5Final',  addr(md5_context));

  name_hashed = put (substr (md5_context,89, 16), $hex32.);

  label name_hashed = 'Name (hashed per MD5)';

  drop md5: name;
run;
filename sascbtbl;

ods listing;
proc print data=work.class;
run;