Good catch, and a bit snappier as well - the below test on 400 clobs at 1.5meg each
SQL> set timing on
SQL> select tag, count(*)
2 from (
3 select x,
4 case when dbms_crypto.hash(y,2) = dbms_crypto.hash(z,2) then 'same' else 'different' end tag
5 from t
6 )
7 group by tag;
TAG COUNT(*)
--------- ----------
different 200
same 200
Elapsed: 00:00:05.95
SQL>
SQL>
SQL> select tag, count(*)
2 from (
3 select
4 x,
5 case when dbms_lob.compare(y, z) = 0 then 'same' else 'different' end tag
6 from t
7 )
8 group by tag;
TAG COUNT(*)
--------- ----------
different 200
same 200
Elapsed: 00:00:03.39
SQL>