Hashes, Checksums and all the rest...

[lights]

I have been using a macro variable with a comma delimited list of variables i want to sent to md5. so it looks like

%let keyvars = %bquote(subjid, lbrefid, lbreqno, lbtestr, lbdat, lbtpt, lbrptid, lbmethod);

to make it easier to live with I use a put function in creating the checksum variable.

data x;

  set y;

  length checksum $32;

  checksum = put(md5(strip(compress(catx(&labkeys),,'c'))), hex20.);

run;

I found it helpful to compress out control characters, forget why at this point

I was using this to create a composite key to be used to join or compare datasets, It has proven quite reliable until I discovered just this week that two datasets may have the same variable names, labels and values but if the length is different in any of the variables the checksum values will be different.  That kind of blew things up.  I was looking around today and came on this post about a new SHA256 function available in SAS 9.4 M1.

http://blogs.sas.com/content/sasdummy/2014/01/18/sha256-function-sas94/