How to XOR md5 hash values and cast them to HEX in postgresql

Question

What I have tried so far

     SELECT md5(text) will return text (hex strings) .

After that We need to xor them

    SELECT x'hex_string'  # x'hex_string';

But the above results in binary values. How do I again convert them into hex string?

Is there anyway to xor md5 values in postgresql and convert this into hexadecimal values again ?

Answer 1

Those binary values are in fact of type bit varying, which differs significantly from bytea.

bit varying comes with built-in support for XOR and such, but PostgreSQL doesn't provide a cast from bit varying to bytea.

You could write a function that does the cast, but it's not trivial and probably not the more efficient way in your case.

It would make more sense to XOR the md5 digests directly. PostgreSQL doesn't provide the XOR operator for bytea either, but it can be easily written in a function, especially when assumed that the operands have an equal length (16 bytes in the case of md5 digests):

CREATE FUNCTION xor_digests(_in1 bytea, _in2 bytea) RETURNS bytea
AS $$
DECLARE
 o int; -- offset
BEGIN
  FOR o IN 0..octet_length(_in1)-1 LOOP
    _in1 := set_byte(_in1, o, get_byte(_in1, o) # get_byte(_in2, o));
  END LOOP;
 RETURN _in1;
END;
$$ language plpgsql;

Now the built-in postgresql md5 function that produces an hex string is not the best fit for post-processing, either. The pgcrypto module provides this function instead:

 digest(data text, type text) returns bytea

Using this function and getting the final result as an hex string:

   select encode(
          xor_digest ( digest('first string', 'md5') ,
                       digest('second string', 'md5')),
          'hex');

produces the result: c1bd61a3c411bc0127c6d7ab1238c4bd with type text.

If pgcrypto can't be installed and only the built-in md5 function is available, you could still combine encode and decode to achieve the result like this:

select
encode(
 xor_digest(
   decode(md5('first string'), 'hex'),
   decode(md5('second string'), 'hex')
 ),
 'hex'
);

Result:

c1bd61a3c411bc0127c6d7ab1238c4bd

Answer 2

You may want to check

https://github.com/artejera/text_xor_agg/

I recently wrote this postgres script for such a use case.