SHA256 Hashing Function in Stata?

[Francois Durant]

Hi,

is there a simple subroutine that would allow to get the output of the SHA256 of a variable in Stata? Thanks.

[Bjarte Aagnes]

You may modify the following:

sysuse auto , clear
keep make
gen str64 SHA256 = ""

java:

import com.stata.sfi.*;

import java.security.MessageDigest;
import java.security.NoSuchAlgorithmException;
import java.nio.charset.StandardCharsets;
import java.math.BigInteger;

MessageDigest md = MessageDigest.getInstance("SHA-256");

String VarName = "make"; 
String VarNameSHA256 = "SHA256"; 

Integer orgstrvar = Data.getVarIndex(VarName);
Integer newstrvar = Data.getVarIndex(VarNameSHA256)
 
for (long n = 1;  n <= Data.getObsTotal(); n++) {

    String text = Data.getStr​(orgstrvar,n);
            
    md.update(text.getBytes(StandardCharsets.UTF_8));
    byte[] digest = md.digest();
    
    System.out.println(String.format("%064x", new BigInteger(1, digest))); 
    Data.storeStrf​(newstrvar, n, String.format("%064x", new BigInteger(1, digest)));
}    
 
end

. list in 1/5

     +----------------------------------------------------------------------------------+
     | make                                                                      SHA256 |
     |----------------------------------------------------------------------------------|
  1. | AMC Concord     43a42426f1ef0ff3852283ee136b6c6175d38187938cea5464853a0362d4a00c |
  2. | AMC Pacer       75e3c7fd9a49b153a270cf8afe6df395c5171775f5dbd642dceefe5b2689eb6f |
  3. | AMC Spirit      b8c9a35beea6d0fdf238947ec5a4c7c063a540d6916a731e76bd2f5c5251c76b |
  4. | Buick Century   1996df3009e4df9a2ed780572723a5d985971961015f980a3186b8cc2081dae9 |
  5. | Buick Electra   b0403019f9f9016e2b008e24bb094563e116b85e081c445c831c9c0f60097b96 |
     +----------------------------------------------------------------------------------+

[Francois Durant]

Thanks a lot Bjarte, it helps!

[Matthew Hall]

If you're using a Unix-based OS (macOS or Linux) you can achieve this using my shell wrapper inshell (available on the SSC) like so:

sysuse auto, clear
gen str64 SHA256 = ""
local var make                  // set your variable here
forvalues i = 1 / `c(N)' {
  inshell echo -n `=`var'[`i']' | shasum -a 256 | tr -d ' -'
  replace SHA256 = r(no1) in `i'
}

. list make SHA256 in 1/5

     +----------------------------------------------------------------------------------+
     | make                                                                      SHA256 |
     |----------------------------------------------------------------------------------|
  1. | AMC Concord     43a42426f1ef0ff3852283ee136b6c6175d38187938cea5464853a0362d4a00c |
  2. | AMC Pacer       75e3c7fd9a49b153a270cf8afe6df395c5171775f5dbd642dceefe5b2689eb6f |
  3. | AMC Spirit      b8c9a35beea6d0fdf238947ec5a4c7c063a540d6916a731e76bd2f5c5251c76b |
  4. | Buick Century   1996df3009e4df9a2ed780572723a5d985971961015f980a3186b8cc2081dae9 |
  5. | Buick Electra   b0403019f9f9016e2b008e24bb094563e116b85e081c445c831c9c0f60097b96 |
     +----------------------------------------------------------------------------------+

Note that you must use the -n flag with echo to remove the trailing new line.